<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Gaurav Kumar on Medium]]></title>
        <description><![CDATA[Stories by Gaurav Kumar on Medium]]></description>
        <link>https://medium.com/@itstheanurag?source=rss-91a5dc5af290------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*lu_iC3yEUe_FGUlTH5sPaw.jpeg</url>
            <title>Stories by Gaurav Kumar on Medium</title>
            <link>https://medium.com/@itstheanurag?source=rss-91a5dc5af290------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 01:05:29 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@itstheanurag/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Week 45 — Products to look out for this week.]]></title>
            <link>https://itstheanurag.medium.com/week-45-products-to-look-out-for-this-week-3c0afe534fe4?source=rss-91a5dc5af290------2</link>
            <guid isPermaLink="false">https://medium.com/p/3c0afe534fe4</guid>
            <category><![CDATA[open-source]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[build-in-public]]></category>
            <category><![CDATA[weekly-product-roundup]]></category>
            <dc:creator><![CDATA[Gaurav Kumar]]></dc:creator>
            <pubDate>Sun, 08 Nov 2026 16:13:10 GMT</pubDate>
            <atom:updated>2026-11-08T16:13:10.047Z</atom:updated>
            <content:encoded><![CDATA[<p>So I didn’t get many products in my tweet this week, but we are here writing the second edition of this series, where every week we talk about a few of the products found on X/Twitter.</p><h4><a href="https://www.remjobs.works/">Remjobs</a></h4><p>Searching for jobs on LinkedIn, Indeed, and Naukri can get pretty boring. You visit a website, enter a few filters, and scroll through hundreds of job listings, most of which you’ll probably never apply to.</p><p><strong>RemJobs takes a slightly different approach.</strong> It feels less like a traditional job hunt and more like exploring the world through Google Earth.</p><p>Imagine having a globe where you can explore startups across different countries, zoom into cities, discover companies, and see which ones are hiring.</p><p>That’s essentially what RemJobs offers.</p><p>The platform currently maps over 253,000 startups across 178 countries. And these aren’t just random company listings. You can explore startups by location, industry, and hiring status, including companies offering remote opportunities.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EJv_cRwecKK_Z-sGG83g_A.png" /></figure><p>According to the job page, it currently has 21k jobs available across 178 countries, so maybe your next job can be found on remojobs.</p><p>What I particularly like is that the platform goes beyond simply showing you job openings. You can discover the companies behind those opportunities, explore their profiles, and even find information about their founders and ways to contact them.</p><h4><a href="https://avaraui.com/">Avara UI</a></h4><p>Let’s be honest, we probably have enough component libraries by now. Every other week, there’s a new collection of buttons, cards, inputs, and other UI elements claiming to make our lives as developers easier.</p><p>So, what makes Avara UI worth checking out?</p><p>Well, instead of focusing only on the usual building blocks, Avara UI puts a lot more emphasis on interactions and animations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OvOrq7UxYoEiR6Afi-ZlwA.png" /></figure><p>Think interactive wallet cards, physics-based falling tags, 3D carousels, animated docks, and other elements that make an interface feel a little more alive.</p><p>And it’s not just about animations. The library also includes complete website sections, from hero sections and navigation bars to pricing tables and dashboards. So you can use it to build an entire page rather than stitching together individual components.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4iPggTMzphBUO6W-C777ww.png" /><figcaption>liked this bento, can’t capture animations in pic, but its good. to Really liked this Bento layout. It&#39;s difficult to capture the animations in a screenshot, but trust me, they look good.</figcaption></figure><p>Another thing I like is how it integrates with the shadcn ecosystem. You can install components directly into your project using the shadcn CLI, modify the source code, and make them fit your own design system. No additional wrapper package to depend on.</p><p>It even supports the shadcn MCP workflow, meaning AI coding assistants can browse and install components from its registry.</p><p>Is it going to replace your existing component library? Probably not. But if you’re building a website and want to add a little more personality through animations and interactions, Avara UI is worth exploring.</p><h4><a href="https://tryspar.dev/">Spar</a></h4><p>We’ve all been there. You open LeetCode, pick a topic, solve a few problems, get stuck on one, look at the solution, and move on.</p><p>Then a week later, you encounter a similar problem and somehow make the same mistakes again.</p><p>The problem isn’t always that you haven’t solved enough questions. Sometimes, you just haven’t figured out what you’re getting wrong.</p><p>That’s the issue <strong>Spar is solving.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*OD5GRPYEKgOFAcg3-zT4pQ.png" /><figcaption>Not gonna lie, the landing page is sick to Not gonna lie, the landing page alone is worth checking out.</figcaption></figure><p>Instead of giving you another list of coding problems to work through, Spar tries to figure out what you actually need to practise.</p><p>It tracks your coding attempts, identifies the concepts you’re struggling with, remembers which test cases break your solutions, and uses all that information to decide what you should work on next.</p><p>Not gonna lie, the landing page is sick</p><h4><a href="https://buildrship.xyz/">BuilderShip</a></h4><p>Submitting a product to a launch platform usually isn’t the most exciting part of building something.</p><p>You open a form, fill in your product name, description, tagline, website URL, upload screenshots, choose categories, and finally hit submit.</p><p>BuildrShip makes this experience feel a little different.</p><p>What I particularly liked was its conversational approach to launching a product. Instead of feeling like you’re filling out yet another submission form, the process feels more like having a conversation with ChatGPT.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ser0tractEUHZlBrXqXAjw.png" /></figure><p>I actually launched <a href="https://rebupl.fun/">Rebupl</a> here, and all I had to do was provide my product URL. The AI took care of the rest, gathering the information and putting together the product listing.</p><p>Gotta be honest, this was probably the easiest product submission I’ve done for Rebupl so far.</p><p>If you’re building something of your own, give BuildrShip a try. You might find launching your product a lot less tedious than you’re used to.</p><p>And if you do, let me know how your experience goes!</p><p>That’s it for this week’s discoveries. Didn’t have a huge list this time, but these were some interesting products I came across while browsing X.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3c0afe534fe4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Is software really dead now? Should you still learn to code?]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://itstheanurag.medium.com/is-software-really-dead-now-should-you-still-learn-to-code-a4a488f0f1fe?source=rss-91a5dc5af290------2"><img src="https://cdn-images-1.medium.com/max/1672/1*LxsiCxmkcRiDwnu72QSRgg.png" width="1672"></a></p><p class="medium-feed-snippet">A few years ago, knowing how to code was a significant advantage.</p><p class="medium-feed-link"><a href="https://itstheanurag.medium.com/is-software-really-dead-now-should-you-still-learn-to-code-a4a488f0f1fe?source=rss-91a5dc5af290------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://itstheanurag.medium.com/is-software-really-dead-now-should-you-still-learn-to-code-a4a488f0f1fe?source=rss-91a5dc5af290------2</link>
            <guid isPermaLink="false">https://medium.com/p/a4a488f0f1fe</guid>
            <category><![CDATA[thoughts-on-coding]]></category>
            <category><![CDATA[ai-driven-development]]></category>
            <category><![CDATA[conversations]]></category>
            <category><![CDATA[coding]]></category>
            <dc:creator><![CDATA[Gaurav Kumar]]></dc:creator>
            <pubDate>Wed, 04 Nov 2026 16:00:19 GMT</pubDate>
            <atom:updated>2026-11-04T16:00:19.400Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[My issues with the Launch directories led me to build a better one.]]></title>
            <link>https://itstheanurag.medium.com/my-issues-with-the-launch-directories-led-me-to-build-a-better-one-9755236ba91c?source=rss-91a5dc5af290------2</link>
            <guid isPermaLink="false">https://medium.com/p/9755236ba91c</guid>
            <category><![CDATA[rebupl]]></category>
            <category><![CDATA[build-in-public]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[launchpad]]></category>
            <category><![CDATA[launch-directory]]></category>
            <dc:creator><![CDATA[Gaurav Kumar]]></dc:creator>
            <pubDate>Tue, 03 Nov 2026 04:35:22 GMT</pubDate>
            <atom:updated>2026-11-03T04:35:22.859Z</atom:updated>
            <content:encoded><![CDATA[<p>Over the past few days, I’ve been exploring several product launch platforms. I wanted to get my newly built product in front of people, so naturally, I started looking at platforms like Product Hunt and other directories where indie developers launch and promote their products.</p><p>But the more platforms I explored, the more frustrated I became.</p><p>Some wanted me to verify ownership before submitting. Others wouldn’t let me launch unless I completed certain tasks on their websites. And then some platforms required me to embed their badges on my own website just to get a listing.</p><p>I understand that these platforms have their own business models and reasons for doing things. But it got me thinking:</p><p><strong>Aren’t these platforms supposed to make product discovery easier for founders? Why does getting listed feel like another marketing task in itself?</strong></p><p>I started wondering whether there was a better way to approach product discovery and what I’d want from a platform if I were building one myself.</p><p>And that’s what this article is about.</p><h4>I am not ungrateful</h4><p>I understand that these platforms solve a real pain point for indie hackers. I also understand that building and maintaining these platforms takes time, effort, and money. They deserve credit for that.</p><p>But goddammit, a platform is supposed to make someone’s life easier, not harder.</p><p>Today, founders aren’t launching on just Product Hunt or one or two other directories. They’re submitting their products to dozens, sometimes hundreds, of launch platforms.</p><p>Now imagine if every single one of them asked founders to embed their badge on their website before allowing them to launch.</p><p>What would your homepage even look like? A wall full of launch badges?</p><p>And how many deployments would you need just to get your product listed everywhere?</p><p>At some point, the entire purpose of these platforms gets lost. Instead of helping founders distribute their products, they’re adding another set of chores to an already long launch checklist.</p><h4>A quick note on dofollow and nofollow links</h4><p>If you’re unfamiliar with these terms, here’s a quick explanation.</p><p>When one website links to another, search engines can use that link as a signal when evaluating the linked website.</p><p>A <strong>dofollow link</strong> is simply a regular hyperlink without a nofollow attribute. It can pass link equity (often called link juice), which may contribute to the linked website&#39;s search engine rankings.</p><p>A <strong>nofollow link</strong> includes the rel=&quot;nofollow&quot; attribute. It tells search engines that the website doesn&#39;t necessarily endorse or want to associate itself with the linked page. Search engines generally treat it as a hint rather than an absolute instruction.</p><p>So, why does this matter?</p><p>Some product directories offer dofollow backlinks as a premium feature, while free listings may receive nofollow links. In other words, founders sometimes have to pay extra to get a link that may provide additional SEO value.</p><h4>Why I built Rebupl</h4><p>That’s where <a href="https://rebupl.fun/">Rebupl</a> comes in.</p><p>As I explored these platforms, I found myself appreciating what I’d built even more.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pLxZjT2-hvahNv7MyA5wNw.png" /></figure><p>Unlike some platforms, I don’t sell dofollow links at different price points. I don’t ask founders to embed badges on their websites just to launch a product. Every product launch gets its own dedicated launch card, even if it’s a free launch.</p><p>I also don’t believe founders should have to wait weeks, let alone months, just to get their products listed.</p><p>Here’s a funny example. I submitted Rebupl to Uneed for a free listing. And guess what launch date I got?</p><p><strong>May 5th.</strong></p><p>Can you believe that? I have to wait more than six months just to get a free listing!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/630/1*h0fyjHN5II65fqQ7SjTZig.png" /></figure><p>I understand that platforms have their own launch schedules and curation processes. But six months? By then, who knows what the product will look like, or whether the launch will even be relevant?</p><p>That’s the kind of friction I wanted to avoid with Rebupl.</p><p>Most indie developers are bootstrapping their businesses, building in their spare time, and trying to make their first dollar. They shouldn’t have to pay multiple platforms just to get noticed. Paid promotions are fine, but basic product discovery should be accessible to everyone, regardless of budget.</p><p>That’s the principle I wanted to build Rebupl around.</p><h4>What exactly do you get on Rebupl?</h4><p>So, what does Rebupl actually offer?</p><p>I wanted to keep things simple. You shouldn’t have to pay just to get your product listed. That’s why Rebupl has a free launch option, and it isn’t just a basic listing with half the features locked away.</p><p><strong>Free — $0, forever</strong></p><p>You can launch up to five products per weekly drop, and every launch gets:</p><ul><li>A permanent, public product page with a dofollow backlink.</li><li>An automatic share card for your launch.</li><li>Community upvotes and leaderboard placement.</li><li>Click and referral analytics.</li><li>Access to 15 free launch tools.</li><li>Unlimited creator enquiries.</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/775/1*mMjTxfYvDuMwtK2q6nsvkw.png" /></figure><p>And yes, the product page stays public and indexable. You don’t have to pay to keep your listing alive.</p><p><strong>Premium Launch — $34.99 (one-time)</strong></p><p>If you want more than a product listing, there’s Premium Launch.</p><p>It includes everything needed for a launch in the next weekly drop, along with an editorial article covering your product, the problem it solves, and who it’s built for.</p><p>The article is also cross-posted to Medium and Substack, giving your product additional exposure. You also get priority placement and a feature in the weekly recap.</p><p>It’s a one-time purchase. No subscriptions, no recurring charges.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/773/1*h_GWiV93V6o-2OoJs8eJxw.png" /></figure><p><strong>Launch Story — $99.99 (one-time)</strong></p><p>This is for founders who want a more dedicated promotional push rather than just a listing or editorial feature.</p><p>It includes everything in Premium Launch, plus:</p><ul><li>A dedicated X post from the Rebupl account.</li><li>A LinkedIn post and newsletter mention.</li><li>One dedicated post in a named community.</li><li>Content written and published by hand within five business days.</li></ul><p>Again, it’s a one-time purchase.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/773/1*7e2TjHdHoR7vgtRSsgPkZg.png" /></figure><p><strong>One thing worth highlighting</strong></p><p>Remember our discussion about dofollow and nofollow links?</p><p>Here’s something I deliberately did differently.</p><p>The free launch actually gets a dofollow backlink. The paid tiers, on the other hand, use sponsored nofollow links.</p><p>Why?</p><p>Because I don’t want to sell link equity as a premium feature. The paid plans are for additional distribution, editorial work, and promotional effort, not for buying a better backlink.</p><p>I think that’s a much fairer approach. Everyone gets a proper product page and a backlink, regardless of whether they spend a single dollar on Rebupl.</p><h4>Rebupl is not only a Launch Pad</h4><p>Another problem founders often face is finding the right people to help get their products in front of a larger audience.</p><p>Getting listed on a launch platform is one thing. Finding creators who can actually help distribute your product is another.</p><p>Keeping that in mind, I added a creator discovery feature to Rebupl.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ysoyNAh-5qQ_GrDJbFu28Q.png" /></figure><p>The idea is simple: why limit distribution to X when there are so many other platforms out there? Instagram, YouTube, LinkedIn, Medium, Substack, and plenty more.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/765/1*jaDzN_URyHttDt1wIIbSOw.png" /></figure><p>Creators can create public pages, list the platforms they work on, and optionally showcase their past work. They can then share these pages with potential clients or get discovered by founders through Rebupl Creators directory.</p><p>You might wonder why a product launchpad needs this feature. Well, I’ve seen indie developers promote other people’s products on X to fund their own projects.</p><p>I wanted Rebupl to make it easier for these creators to connect with founders who need help with distribution.</p><p>After all, it shouldn’t just help founders get discovered; it should help them find the right people to grow their products too.</p><h4>Can I keep this running if the product never makes any money?</h4><p>Honestly, I think so.</p><p>At least for now, I’m running Rebupl on Cloudflare’s free tier. And looking at the kind of traffic and usage I’m expecting initially, I don’t see myself hitting those limits anytime soon.</p><p>Cloudflare gives you 100,000 Worker requests per day on its free plan, and if you’re using D1, there’s a 500 MB limit per database. That’s plenty of room for a small product directory like Rebupl, at least at its current stage.</p><p>And there’s another thing: I don’t necessarily need hundreds of paid launches to keep this going.</p><p>If even one person purchases an editorial feature or a promotional package occasionally, that should be enough to help cover the infrastructure costs.</p><p>Of course, that might change as the platform grows. More users, more traffic, and more data will eventually mean higher infrastructure costs. But that’s a problem I’d be happy to have.</p><p>For now, I don’t want to overcomplicate things or introduce subscriptions just to cover costs that I don’t even have yet.</p><p>I’d rather keep the platform accessible, see how people use it, and figure out monetization as it grows.</p><blockquote>The goal isn’t just to make money. The goal is to provide value, because providing value is what makes you irreplaceable.</blockquote><p>~ <a href="https://medium.com/u/91a5dc5af290">Gaurav Kumar</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9755236ba91c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to write Clean Code…]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-image"><a href="https://itstheanurag.medium.com/how-to-write-clean-code-acc05bab90ff?source=rss-91a5dc5af290------2"><img src="https://cdn-images-1.medium.com/max/600/0*Vz-bO8zCW8nXINhd.jpg" width="600"></a></p><p class="medium-feed-snippet">They say every developer should read Clean Code at least once. It&#x2019;s practically the bible for developers who care about writing good&#x2026;</p><p class="medium-feed-link"><a href="https://itstheanurag.medium.com/how-to-write-clean-code-acc05bab90ff?source=rss-91a5dc5af290------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://itstheanurag.medium.com/how-to-write-clean-code-acc05bab90ff?source=rss-91a5dc5af290------2</link>
            <guid isPermaLink="false">https://medium.com/p/acc05bab90ff</guid>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[how-to-code]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[clean-code]]></category>
            <category><![CDATA[rebupl]]></category>
            <dc:creator><![CDATA[Gaurav Kumar]]></dc:creator>
            <pubDate>Mon, 02 Nov 2026 14:07:37 GMT</pubDate>
            <atom:updated>2026-11-02T14:07:37.325Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[Week 44 — Product to check out this week]]></title>
            <link>https://itstheanurag.medium.com/week-44-product-to-check-out-this-week-082e1f4c8b4c?source=rss-91a5dc5af290------2</link>
            <guid isPermaLink="false">https://medium.com/p/082e1f4c8b4c</guid>
            <category><![CDATA[launchpad]]></category>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[build-in-public]]></category>
            <category><![CDATA[directory]]></category>
            <category><![CDATA[saas]]></category>
            <dc:creator><![CDATA[Gaurav Kumar]]></dc:creator>
            <pubDate>Sun, 01 Nov 2026 13:55:54 GMT</pubDate>
            <atom:updated>2026-11-01T13:55:54.909Z</atom:updated>
            <content:encoded><![CDATA[<p>I have started this new series, where every Sunday I will publish a blog about the tech product I found. This is the first issue of this series; buckle up.</p><blockquote>i did not take any money from any of these product owners to write this article.</blockquote><h4><a href="https://devatlas.site/">01 — Devatlas</a></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*F1AgirZY4C7MXfzAySbZpg.png" /></figure><p>DevAtlas is a curated directory of tools you need to build and launch your product. From authentication and infrastructure to databases, observability, and payments, it brings essential developer tools together in one place.</p><p>Not sure which tech stack to choose for your next project? DevAtlas has you covered with its <strong>prebuilt stacks</strong> — carefully put-together combinations of tools that help you find the right technologies for your specific needs, so you can spend less time researching and more time building.</p><p>Created and managed by <a href="https://x.com/sankitdev">Sankit</a></p><h4><a href="https://audanweb.xyz/#work">02 — AudanWeb</a></h4><p>Building your product is one thing; making it come across to people is another thing. AudanWeb does just that for you.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0dSn5IKVNq-N3ZObZffpew.png" /></figure><p>Audan Web lets you create campaigns for your product launch and connect with the right creators on X (Twitter) to help you boost your launch posts. They have over 100+ creators; total reach comes to around 500k+ people.</p><p>Their last campaigns made it to X news, so if you are looking to launch your product, check them out.</p><p>Created and managed by <a href="https://x.com/jatinprajapat29">Jatin</a></p><h4><a href="https://www.goatmcp.xyz/">03 — GoatMCP</a></h4><p>In the age of AI, people are relying more and more on AI agents to do the work; these AI agents rely heavily on tools, MCP servers, and skills to accomplish the given tasks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hZp5Esnbo97OQML-6-rmbg.png" /></figure><p>GoatMCP makes that easier by bringing them together in one searchable directory. You can explore MCP servers, agent skills, and agents across categories like frontend, backend, developer tools, AI/ML, design, RAG, video editing, and SEO.</p><p>What makes it particularly interesting is its promise of discovering and installing MCP servers and skills with a single command. Whether you’re using AI coding assistants or building your own agents, GoatMCP gives you a convenient place to discover tools that can extend their capabilities.</p><p>Created and Managed by <a href="https://x.com/hiteshxdev">Hitesh</a></p><h4><a href="https://designeer.xyz/">04 — Designeer</a></h4><p>Unlike DevAtlas, which mostly focuses on the backend side of the app, Designeer deals with the frontend side of things.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lRo6caN2ZZ2p_LHvlQ1haQ.png" /></figure><p>Ever got stuck on a project? Lacking some design inspiration? Well, Designeer is the place you get to inspire yourself; it’s a curated discovery platform for designers, frontend developers, and creative builders.</p><blockquote>Designeer is more than a directory of design tools. I’d describe it as a curated library of inspiration, resources, and tools for people who build for the web.</blockquote><p>Created and Managed by <a href="https://x.com/dhruvtwt_">Dhruv</a></p><h4><a href="https://hunchkey.com/">05 — HunchKey</a></h4><p>Ever find yourself staring at an empty reply box, wondering what to say next? HunchKey is an AI-powered desktop writing assistant that helps you write messages, emails, and complete sentences in your own voice.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*O8nT80Sj1pG-lixqVAWmVA.png" /></figure><p>Beyond generating replies, HunchKey can complete sentences as you type, fix grammatical mistakes, and remember custom spellings for words and names.</p><p>But what makes it particularly interesting is its ability to understand the context of your conversations. It remembers who you’re talking to and adapts your writing persona accordingly.</p><p>Talking to a client? It switches to a more professional tone. Messaging a friend? It keeps things casual and natural. The idea is to make AI-assisted writing feel less like AI and more like you, adjusting to the person on the other end.</p><p>Created and Managed by <a href="https://x.com/preethamrangu">Preetham</a></p><h4><a href="https://rebupl.fun/">06 — Rebupl</a></h4><p>Rebupl is a Product launch and Creator discovery platform. Most people launch on Product Hunt or Peerlist, get a handful of upvotes, share the link on X, and call it a day. But what happens after that?</p><p>Rebupl takes a different approach. It brings product launches and creator marketing together, giving you more ways to get your product in front of the right audience.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pLxZjT2-hvahNv7MyA5wNw.png" /></figure><p><em>Honesty time:</em> Rebupl is actually something I’m building myself. You can explore the platform and use most of its features right now, although payment functionality is still being finalised.</p><p>I’d love for you to check it out, give it a try, and let me know what you think. And if you’d like to follow along as I build, consider subscribing to the newsletter!</p><p>Being built by <a href="https://x.com/Itstheanurag">Gaurav</a></p><h4><a href="https://appfinder.tech/">07 — AppFinder</a></h4><p>Finding the right software for your needs can be surprisingly frustrating. You usually end up opening a dozen tabs, comparing pricing pages, reading reviews, and still wondering which tool is actually worth using.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NlK7Gp5GrbmthUyGY4Ny9w.png" /></figure><p>AppFinder aims to make that process easier. It’s a software discovery platform where you can explore apps across categories like productivity, developer tools, marketing, AI, analytics, and more.</p><p>Instead of knowing exactly which product you’re looking for, you can simply describe what you need and discover relevant options. You can also compare different apps, explore alternatives, and see how different tools relate to each other through its interactive app map.</p><p>Created and Managed by <a href="https://x.com/kunal_twts">Kunal</a></p><h4><a href="https://startupslab.site/">08 — StartupLabs</a></h4><p>StartupLabs is another product launch platform, but I quite like how they’ve approached it. Instead of throwing hundreds of products into one launch day, they have a weekly launch cycle where new products go live every Monday.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Z2KV0BtV8l-_APeeHzaA7g.png" /></figure><p>But there’s more to it than just launching your product.</p><p>They also have a list of 100+ startup directories where you can submit your product (which is pretty useful if you’re trying to get your first few backlinks). They even offer some free tools to help founders prepare their websites for launch and take care of basic SEO stuff.</p><p>If you’re building something and looking for more places to get it in front of people, this is one I’d add to my bookmarks.</p><p>Created and Maintained by <a href="https://x.com/GohilHardy">Hardik</a></p><h4><a href="https://lycheedata.com/">09 — Lychee</a></h4><p>If you follow prediction markets like Kalshi or Polymarket, Lychee is worth checking out.</p><p>It’s a data research and analytics platform that brings live and historical market data into one workspace. You can explore trades, compare odds across platforms, investigate wallets, and track market activity without having to deal with APIs or write SQL queries.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0iVPl-yP6AfMZ0CmW1SQHw.png" /></figure><p>You can also build your own charts, dashboards, and live research terminals, combine data from different sources like Binance and Chainlink, and export datasets for further analysis.</p><p>Basically, it gives you the tools to dig deeper into prediction market data without having to build your own data pipelines or analytics setup.</p><p>Created and Maintained by <a href="https://x.com/misterrpink1">Kash</a></p><p>That’s all for today. If you guys want me to write about your projects, you can reach out to me on <a href="https://x.com/Itstheanurag">X at Gaurav</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=082e1f4c8b4c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Introducing Codecript: Your Software Engineering Learning Hub]]></title>
            <link>https://itstheanurag.medium.com/introducing-codecript-your-software-engineering-learning-hub-4f844e564322?source=rss-91a5dc5af290------2</link>
            <guid isPermaLink="false">https://medium.com/p/4f844e564322</guid>
            <category><![CDATA[codecript]]></category>
            <category><![CDATA[learn-engineering]]></category>
            <category><![CDATA[open-source]]></category>
            <category><![CDATA[rebupl]]></category>
            <category><![CDATA[coding]]></category>
            <dc:creator><![CDATA[Gaurav Kumar]]></dc:creator>
            <pubDate>Sun, 01 Nov 2026 03:39:41 GMT</pubDate>
            <atom:updated>2026-11-01T03:39:41.938Z</atom:updated>
            <content:encoded><![CDATA[<p>In the age of AI, a struggling job market and mass layoffs are slowly becoming the new norm. People go out for lunch and come back to find that their system access has been revoked.</p><p>It’s probably safe to say that the days of getting by with mediocrity are long gone. If you want to stay relevant as a software engineer, you need to keep learning and stay ahead of the curve.</p><p>And it’s not just about solving LeetCode questions anymore.</p><p>With AI making code increasingly easier to produce, simply knowing how to write code isn’t enough. You need to understand what happens under the hood, know your fundamentals, understand system design, databases, networking, operating systems, and much more.</p><p>But how do you prepare for all of this?</p><p>I wanted a single platform where I could practice coding, learn engineering concepts, track my progress, and revisit things I tend to forget.</p><p>Because when you’re constantly writing code, it’s surprisingly easy to forget the theory behind it. You might know how to implement something but struggle to explain why it works.</p><p>I couldn’t find quite what I was looking for, so I decided to build it myself.</p><p><strong>That’s how Codecript came into existence.</strong></p><h4>INTRODUCING CODECRIPT</h4><p>People often underestimate the importance of having a collection of resources, so I built <a href="http://codecript.pages.dev/">Codecript</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-Fo8y9yozxDZ489QkgTxng.jpeg" /><figcaption>codecript</figcaption></figure><p>Codecript is a collection of what you need to prepare for your interviews; whatever you think a software engineer needs ( apart from AI; currently, no AI), we have it here.</p><p>May it be languages, may it be fundamentals, system design concepts, low-level design, high-level design, DevOps concepts, testing- we&#39;ve got you covered under one hood</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*_r-wg8z7ijEgyONx5v1A3g.jpeg" /><figcaption>golang</figcaption></figure><h4>SQL PLAYGROUND</h4><p>If you’re a software engineer, you really can’t afford to ignore SQL.</p><p>Sure, you might not need to master it to build a regular web application, but that’s hardly a reason not to learn it. Sooner or later, you’ll find yourself dealing with databases, writing complex queries, or trying to figure out why a particular query isn’t returning what you expected.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5wIGl_TElGdMLuJAgSd4-w.jpeg" /></figure><p>That’s why I built an SQL Playground in Codecript.</p><p>It runs entirely in your browser, so there’s nothing to install or configure. You can jump straight into practicing SQL, starting with basic SELECT queries and gradually moving towards aggregations, joins, window functions, CTEs, and much more.</p><p>There’s a collection of practice questions covering different SQL concepts, giving you a way to learn by actually writing and executing queries rather than just reading about them.</p><p>And the best part? <strong>No login required.</strong></p><p>Just open the playground, start solving questions, and your previously solved questions stay saved, so you can pick up right where you left off.</p><p>No setup, no unnecessary friction. Just SQL practice. <a href="https://codecript.pages.dev/playground">playground</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tBwmombT-oHzuKr8MrNgig.jpeg" /></figure><p>All concepts practiced in 64 questions.</p><h4>ENGINEERING</h4><p>With AI making code increasingly cheap to produce, the value of an engineer is shifting towards something more fundamental: knowing what to build, how to build it, and why.</p><p>Writing code is only one part of software engineering. Designing systems, making architectural decisions, understanding trade-offs, and building something that can actually serve users are equally important.</p><p>And with that in mind, I wanted Codecript to go beyond conventional coding practice.</p><p>The Engineering section covers different aspects of real-world software engineering, including:</p><ul><li><strong>API Design:</strong> Learn how to design APIs that are consistent, maintainable, and scalable.</li><li><strong>System Design:</strong> Understand how to architect applications, handle scale, and make the right engineering trade-offs.</li><li><strong>Low-Level Design:</strong> Get into the details of object-oriented design, design patterns, and writing maintainable code.</li><li><strong>Building:</strong> Step-by-step guides to building actual software projects, understanding the decisions involved at every stage.</li></ul><p>The goal isn’t just to help you prepare for your next interview. It’s to help you become a better engineer, someone who can take an idea, make the right architectural decisions, and turn it into working software.</p><p>Because knowing how to solve a problem is important, but knowing how to build the entire system around that solution is a different skill altogether.</p><h4>Closing with the truth</h4><p>I’m building Codecript alone, so there’s only so much I can do at a time. There are probably things I’ve missed, features that could be improved, or ideas that haven’t crossed my mind yet.</p><p>I’d genuinely love to hear your feedback. Whether it’s a bug, a missing feature, or simply an idea that could make Codecript better, feel free to share it.</p><p>And if you find the project useful, consider giving it a star on GitHub. It might seem like a small thing, but it goes a long way for an independent builder.</p><p><strong>Check out Codecript:</strong> <a href="https://codecript.pages.dev/">https://codecript.pages.dev</a></p><p><strong>GitHub:</strong> <a href="https://github.com/itstheanurag/codecript">https://github.com/itstheanurag/codecript</a></p><p>Thanks for reading. And happy learning!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4f844e564322" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Building a launch directory that rents out People.]]></title>
            <link>https://itstheanurag.medium.com/building-a-launch-directory-that-rents-out-people-588ed2956d03?source=rss-91a5dc5af290------2</link>
            <guid isPermaLink="false">https://medium.com/p/588ed2956d03</guid>
            <category><![CDATA[build-in-public]]></category>
            <category><![CDATA[saas]]></category>
            <category><![CDATA[side-project]]></category>
            <category><![CDATA[directory-saas]]></category>
            <category><![CDATA[nextjs]]></category>
            <dc:creator><![CDATA[Gaurav Kumar]]></dc:creator>
            <pubDate>Fri, 09 Oct 2026 15:45:12 GMT</pubDate>
            <atom:updated>2026-10-09T16:05:02.368Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*USQ7F5_RMBHT_93-.png" /></figure><p>Lately, I’ve noticed people making good money from their products, especially in the Product Hunt alternatives space. This inspires others to launch their own products. It’s a fascinating space that naturally attracts people, and for serious creators, it can be a goldmine.</p><p>With just a click of a button, creators can discover what their competitors are building and identify products that are gaining significant traction and attracting thousands of views.</p><p>That got me thinking: why not build a launchpad of my own, but with much greater earning potential?</p><p>Any serious indie developer would love to have their product promoted by people who know how to do it well.</p><p>X (formerly known as Twitter) is one such platform where paid promotional campaigns are common. One of my mutuals made over $1,000 just by promoting products, despite having fewer than 5,000 followers on X.</p><p>This got me thinking: what if there were a launchpad that didn’t just help indie developers launch their products, but also helped them get those products in front of the right audience?</p><p>Imagine connecting developers with creators who already know how to promote products, have an engaged audience, and can help turn a launch into actual customers.</p><p>There’s clearly an opportunity here, but I think there’s more to it than just building another Product Hunt alternative.</p><p>I’m still exploring the idea, but I want to understand how this space works, where the existing platforms fall short, and whether there’s a better way to connect product launches with creator-led promotion.</p><h4>So, what am I building?</h4><p>This article is more or less a reflection of my thoughts as I work on this idea. To be completely honest, I know very little about this space. Right now, all I have is an idea, a domain, and a bare-minimum MVP running at <a href="https://rebupl.fun/">rebupl.fun</a>.</p><p>Since I’m new to this space, my first step was to explore existing platforms. I wanted to understand what they offer, how they work, and, more importantly, what I could do differently.</p><p>I’m not trying to build just another launchpad with a slightly different design. I want to figure out whether there’s a genuinely better way to help indie developers get their products noticed while allowing creators to earn by promoting products they believe in.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*qba6_xfCatF5KJj8.png" /></figure><p>That’s what I’m hoping to explore as I build Rebupl.</p><h4>The Stack</h4><p>I’m primarily a backend developer. I’ve built frontend applications with the help of AI from time to time, but I’ve never really tried to turn one into a business. When it comes to design, let’s just say my taste isn’t something I’d confidently bet on.</p><p>Since Rebupl isn’t a particularly complex project, I decided to keep the tech stack simple.</p><p>I’m using <strong>Next.js</strong> for the application and hosting it on <strong>Cloudflare Workers</strong>. I wanted to avoid the hassle of setting up and maintaining email infrastructure for authentication, so I went with Google OAuth. No need to build ten different endpoints and five different screens just to let someone log in.</p><p>In a way, I’m following the good old KISS principle: <em>Keep It Simple, Stupid.</em></p><p>For the database, I’m using <strong>Cloudflare D1</strong>, which is based on SQLite. This is my first time working with SQLite. So far, I’ve mostly worked with PostgreSQL and MongoDB, so it’ll be interesting to see how things differ.</p><p>I’m using <strong>Drizzle ORM</strong> to interact with the database. That should make writing queries more convenient, although I know it won’t magically eliminate the differences between SQLite and PostgreSQL. Things like database features, supported data types, and query behaviour can still differ.</p><p>For now, this stack seems like a reasonable choice for a relatively small project. I don’t want to overengineer something that hasn’t even found its users yet.</p><p>I’d rather spend my time figuring out what people actually want from Rebupl than building infrastructure for problems I might never have.</p><h4>The Payments</h4><p>I’ve been seeing a lot of people building things on X, and Dodo Payments seems to be a popular choice among them. I know it’s a pretty silly reason to choose a payment platform just because I keep seeing people use it, but that’s exactly what happened. So, Dodo Payments it was.</p><p>Before this, I’d worked with a payment gateway called Square. If you haven’t heard of it, don’t worry; it just doesn’t seem to get talked about much in the circles I follow.</p><p>Square was pretty straightforward to work with. You created your plans, set up subscriptions, configured a webhook endpoint, and started listening for events. Simple enough.</p><p>There was one thing that bothered me, though. I couldn’t find a webhook signing secret in the Square setup I was working with. I knew that verifying webhook signatures was important for security, but the implementation I encountered didn’t seem to have a clear way of doing it. At the time, the client was set on using Square, so I worked with what I had.</p><p>Dodo Payments turned out to be a different experience.</p><p>The webhook signing secret was there, and understanding how it was used to verify incoming webhook requests was an interesting part of the integration. The payment flow itself was also quite different. Creating products, handling payments on the frontend, and connecting everything to the backend required a different way of thinking.</p><p>I haven’t manually written every part of the integration from scratch, but I did spend time reading through the code to understand how everything worked. And honestly, that was one of the more interesting parts of the process.</p><p>Payments are one of the most important parts of any website, so naturally I have to test it more and more, which I keep doing.</p><h4>The End Goal</h4><p>For all I know, this product might turn out to be a dud. Nobody might ever use it, and that’s a very real possibility.</p><p>Even so, this project matters to me because I’ve never built a SaaS product before.</p><p>Maybe the first one takes me two or three weeks to build, even with AI helping me along the way. But I’m hoping that repeating this process will gradually bring that time down to a matter of days. The faster I can build, launch, learn, and iterate, the better.</p><p>Failing fast and learning from it is far more valuable than spending months chasing an idea that might never work.</p><p>Of course, that’s easy to say when you’re still writing code. The real challenge begins when you have to put your product out there and convince people to actually use it.</p><p>Since I’ve already gone through the trouble of buying a domain, writing the code, and getting the damn thing deployed, I might as well see this through.</p><p>Now comes the part I’ve been putting off: marketing.</p><h4>Marketing</h4><p>How do I market a product when I’m not particularly good at convincing people to do anything?</p><p>Well, I’m not entirely sure yet. But I do know one thing: I can’t just build a product, put it on the internet, and expect people to magically discover it.</p><p>I’ll have to reach out to people first. Tell them about Rebupl, ask them to give it a try, and see what they think. Maybe I’ll post about it in a few relevant Reddit communities, start sharing my journey on X, and publish an article or two on Substack (including this one).</p><p>I don’t know which of these will work, or whether any of them will work at all. But I suppose that’s the point. I need to try different things, see what gets a response, and double down on whatever seems to be working.</p><p>I’m not particularly comfortable with marketing, and I’m certainly not an expert at selling things. But if I want to build products that people actually use, I need to learn how to get them in front of the right people.</p><p>For now, that’s the plan. Nothing fancy, no grand marketing strategy. Just putting myself and my product out there, talking to people, and figuring things out as I go.</p><p>Let’s see how badly I can mess this up.</p><p><em>Originally published at </em><a href="https://substack.com/home/post/p-219568525"><em>https://itstheanurag.substack.com</em></a><em> on October 9, 2026.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=588ed2956d03" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Content Creation is more than just views and likes.]]></title>
            <link>https://itstheanurag.medium.com/content-creation-is-more-than-just-views-and-likes-81db68395b5b?source=rss-91a5dc5af290------2</link>
            <guid isPermaLink="false">https://medium.com/p/81db68395b5b</guid>
            <category><![CDATA[twitter]]></category>
            <category><![CDATA[content-creation]]></category>
            <category><![CDATA[instagram]]></category>
            <category><![CDATA[youtube]]></category>
            <dc:creator><![CDATA[Gaurav Kumar]]></dc:creator>
            <pubDate>Wed, 24 Jun 2026 07:59:51 GMT</pubDate>
            <atom:updated>2026-06-24T07:59:51.196Z</atom:updated>
            <content:encoded><![CDATA[<p>Recently, I’ve been trying to create content (some people might call it “slop”), and the process has taught me a few things that I think are worth sharing.</p><p>When we talk about creating something, most people focus on the final output: the video, the article, the tweet, or the post. But creation is much more than making something and putting it in front of an audience. The interesting part is the thought process behind it.</p><p><strong>Why Create Anything?</strong></p><p>One realization I had is that not all content is meant for everyone.</p><p>We often assume that every piece of content is trying to inform, entertain, or persuade us. In reality, different creators have different motivations. Some create content simply to provoke a reaction and capture attention. Others create because they want to teach, document, or inspire.</p><p>Sometimes the intention isn’t even conscious. A person sharing their work, their progress, or their daily workout may not be trying to motivate anyone, yet someone else sees it and feels inspired to start their own journey.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/750/1*ud8S7F54RsGmgSMOzqYqKg.png" /></figure><p>This is a quote that has been overused. Here’s an idea: create proof that you tried. Writing articles and creating videos, or recording your process, is a way for you to create proof that you did your best.</p><p><strong>Preparing for content beyond skills</strong></p><p>From my own experience, there are pre-content struggles and post-content struggles that one needs to prepare to face.</p><p><strong>Pre-content struggle</strong></p><p>See here’s the thing, scrolling reels, X, or any other social media for hours is easy, picking up the camera and talking about something for a minute or ten is hard, reading an article is easy, writing one is hard. not cause of the work that it involves, but because every post opens you up to criticism, every idea that you share online can and will get challenged, and every comment will not be in favour of it. knowing all that and still doing it, something impressive in its own way. That’s why not everyone can do it.</p><p><strong>Post-content struggle</strong></p><p>There’s this famous quote from Mike Tyson</p><blockquote><em>“Social media made y’all disrespecting people and not getting punched in the face for it”</em></blockquote><p>Ironically, this holds for many people, including me. Every post that I make or someone else makes opens them to criticism. Every gym video that I share has people complaining about something. Every meal pic that I share, someone has a problem with either protein, carbs, or fiber. There are only two realistic ways to deal with criticism.</p><ul><li>Take them as constructive, but not from the people who do not create content in your genre, or have never created or shared anything in that genre.</li><li>Say “you are right, I need to do better,” and then forget about it. Not all opinions are worth getting triggered over; some simply need to be heard and forgotten about.</li></ul><p>If you can’t do either of those, then you are not mature enough to create anything.</p><p><strong>Never-ending chase of views that consumes you</strong></p><p>Recently, a tech creator decided to “kill” every major technology known to mankind. The actual ideas in the series aside, the titles and thumbnails were clearly engineered to grab attention.</p><p>And honestly, there’s nothing wrong with that.</p><p>People watched the videos. I watched them too. If content creation is how you pay your bills, then chasing views is part of the job. Attention is the currency of the internet, and creators have every right to compete for it.</p><p>The problem begins when views stop being a metric and become your identity.</p><p>If every like excites you and every low-performing post ruins your day, you’re no longer creating content — you’ve become dependent on its validation.</p><p>At that point, your happiness is being outsourced to an algorithm.</p><p>Every waking moment becomes a question of numbers. How many views? How many likes? Why did that post perform better? Why did this one fail?</p><p>The chase never ends because there is no finish line. The video that gets 1,000 views makes you want 10,000. The one that gets 10,000 makes you want 100,000.</p><p>The goal keeps moving.</p><p>If you find yourself obsessing over metrics, close the analytics dashboard and go do something else. Read a book. Go for a walk. Touch grass, as the internet likes to say.</p><p>Because if you’re not careful, content creation can slowly consume your entire personality.</p><p>And that’s no way to live.</p><p><strong>Why do I create content even when I suck at it?</strong></p><p>There was a time when I got almost no views on X. The payout I received recently was my fifth consecutive payout cycle, but for a long time, there was nothing.</p><p>There was a time when my Instagram analytics would show a few hundred views for an entire month. Today, it’s closer to 30,000 views in the last seven days. I still suck at it, but that’s not the point.</p><p>There was a time when I would go to the gym, finish my workout, and move on with my day. Now I record some of those workouts and share them on X, Instagram, and YouTube. People have started following me.</p><p>There was also a time when I had a lot to say about things and simply didn’t. I’d type out a post, read it once, and hit the delete button.</p><p>Looking back, I think content creation taught me something simple: I can do more than I think.</p><p>I can share ideas that I would have kept to myself. I can document things I would have forgotten. I can put my work out into the world and let people decide what they think of it.</p><p>Naturally, I’m not great at it. Most of what I create will probably never go viral. Most of it won’t even be remembered.</p><p>But that’s okay.</p><p>I don’t need every post to be a win. I only need a few wins that matter.</p><p>And the only way to find those few wins is to keep creating.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=81db68395b5b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Stack and Queue in JavaScript: A Node-Based Implementation]]></title>
            <link>https://itstheanurag.medium.com/stack-and-queue-in-javascript-a-node-based-implementation-c74796a5591f?source=rss-91a5dc5af290------2</link>
            <guid isPermaLink="false">https://medium.com/p/c74796a5591f</guid>
            <category><![CDATA[data-structures]]></category>
            <category><![CDATA[stack]]></category>
            <category><![CDATA[sd]]></category>
            <category><![CDATA[queue]]></category>
            <dc:creator><![CDATA[Gaurav Kumar]]></dc:creator>
            <pubDate>Thu, 19 Mar 2026 14:54:36 GMT</pubDate>
            <atom:updated>2026-03-19T14:54:36.715Z</atom:updated>
            <content:encoded><![CDATA[<p>The best way to prepare for interviews is to practice data structures and algorithms in tech, and the best way to learn Data Structures is to build them from scratch.</p><p>Which brings us to the title of this article. Although a lot of you are familiar with them, some still might need this.</p><h4>Stack</h4><p>A <strong>stack</strong> is a linear data structure based on the <strong>LIFO (Last-In, First-Out)</strong> principle. meaning the element that is added last will be removed first.</p><p>Stacks are commonly used to solve complex problems, such as:</p><ul><li>Function calls</li><li>Undo/redo operations</li><li>Expression evaluation</li><li>Backtracking algorithms</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zF3M_f4jwWSm3FSyy1EgQQ.png" /><figcaption>an image created using tldraw canvas</figcaption></figure><p>In JavaScript, the concept of a stack appears in the form of the <strong>call stack</strong>. The JavaScript engine uses a stack to manage execution contexts:</p><ul><li>The <strong>Global Execution Context</strong> is pushed first (at the bottom of the stack)</li><li>Every time a function is invoked, a new <strong>Execution Context</strong> is pushed on top</li><li>When a function finishes execution, its context is <strong>popped off the stack</strong></li></ul><p><strong><em>Implementation</em></strong></p><p>For this implementation, we will use a <strong>node-based approach</strong> instead of an array.</p><p>While arrays can be used for a simple implementation, using a <strong>class-based node structure</strong> gives us better control over the data structure and helps us understand how stacks work under the hood.</p><pre>// basic node class<br>class Node {<br>  constructor(value) {<br>    this.value = value;<br>    this.next = null;<br>  }<br>}<br><br>// a stack class<br>class Stack {<br>  constructor() {<br>    this.top = null; // Reference to the top of the stack<br>    this.bottom = null; // Reference to the bottom of the stack<br>    this.size = 0; // Number of elements in the stack<br>  }<br><br>  /**<br>   * Adds an element to the top of the stack.<br>   * @param {*} value - The value to be added<br>   * @returns {number} The new size of the stack<br>   */<br>  push(value) {}<br><br>  /**<br>   * Removes and returns the element from the top of the stack.<br>   * @returns {*|null} The value of the removed node, or null if stack is empty<br>   */<br>  pop() {}<br>}</pre><p><strong><em>Base Classes</em></strong></p><p>We start by defining our base classes.</p><blockquote><em>Note: The </em><em>Node class will remain the same for both </em><strong><em>Stack</em></strong><em> and </em><strong><em>Queue</em></strong><em> data structures.</em></blockquote><p>The stack itself maintains three properties:</p><ul><li>top → points to the top element of the stack</li><li>bottom → points to the bottom element</li><li>size → keeps track of the number of elements</li></ul><p><strong>Stack Operations</strong></p><p>The stack will have two primary methods:</p><ul><li><strong>push(value)</strong> → adds a new element to the <strong>top</strong> of the stack</li><li><strong>pop()</strong> → removes the element from the <strong>top</strong> of the stack</li></ul><p>These operations ensure the <strong>LIFO (Last-In, First-Out)</strong> behavior.</p><h4>Code</h4><pre>class Node {<br>  constructor(value) {<br>    this.value = value;<br>    this.next = null;<br>  }<br>}<br><br>class Stack {<br>  constructor() {<br>    this.top = null; // Reference to the top of the stack<br>    this.bottom = null; // Reference to the bottom of the stack<br>    this.size = 0; // Number of elements in the stack<br>  }<br><br>  push(value) {<br>    const node = new Node(value);<br>    // If stack is empty, set both top and bottom to the new node<br>    if (!this.top) {<br>      this.top = node;<br>      this.bottom = node;<br>    } else {<br>      // Link the new node to the current top and update top<br>      const temp = this.top;<br>      this.top = node;<br>      this.top.next = temp;<br>    }<br><br>    return this.size++;<br>  }<br><br>  /**<br>   * Removes and returns the element from the top of the stack.<br>   * @returns {*|null} The value of the removed node, or null if stack is empty<br>   */<br>  pop() {<br>    if (!this.top) return null;<br><br>    let temp_node = this.top;<br><br>    // If there&#39;s only one element, clear the bottom reference<br>    if (this.top === this.bottom) this.bottom = null;<br><br>    // Move top pointer to the next node<br>    this.top = this.top.next;<br>    this.size--;<br><br>    return temp_node.value ?? null;<br>  }<br>}</pre><p>The push method adds a new element to the top of the stack. It creates a new node and updates the top pointer to point to it. If the stack is empty, both top and bottom are set to this node. Otherwise, the new node simply points to the current top, and then becomes the new top. Since this only involves a few pointer updates and no traversal, the operation runs in <strong>O(1)</strong> time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*QXKXRJd4_5XbMPb7Hu1wag.png" /><figcaption>push method</figcaption></figure><p>Thepop method removes the element from the top of the stack and returns its value. If the stack is empty, it returns null. Otherwise, it updates the top pointer to the next node and reduces the size. If there were only one element in the stack, both top bottom would be reset null after removal. Likepush, this operation also runs in <strong>O(1)</strong> time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*a5bu6sugHTTnUs51utX22w.png" /><figcaption>push and pop operations</figcaption></figure><p>Even though thebottom pointer isn’t strictly required for a stack; keeping it updated helps maintain a complete and consistent structure.</p><h4><strong><em>Queue</em></strong></h4><p>A <strong>queue</strong> is a linear data structure based on the <strong>FIFO (First-In, First-Out)</strong> principle. This means the element that is added first will be the first one to be removed.</p><p>You can think of it like a line of people — the person who arrives first gets served first.</p><p>Queues show up everywhere in systems:</p><ul><li>Handling requests in a server</li><li>Task scheduling</li><li>Message queues (Kafka, RabbitMQ)</li><li>Print jobs</li></ul><p>Whenever order matters and fairness is required, queues are used.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NO9Ro7fnonttFfUmEflfDw.png" /></figure><p>Just like in the stack implementation, we will use the same Node class to represent each element in the queue.</p><p>The queue itself maintains two main references:</p><ul><li>first → points to the first element in the queue (the one to be removed)</li><li>last → points to the last element (where new elements are added)</li></ul><h4>Code</h4><pre><br>/**<br> * Queue is a FIFO (First In, First Out) data structure.<br> * Elements are added at the rear and removed from the front.<br> */<br>class Queue {<br>  constructor() {<br>    this.first = null;<br>    this.last = null;<br>    this.size = 0; <br>  }<br><br>  /**<br>   * Adds an element to the rear of the queue.<br>   * @param {*} value - The value to be added<br>   * @returns {number} The new size of the queue<br>   */<br>  enqueue(value) {<br>    const node = new Node(value);<br><br>    // If queue is empty, set both first and last to the new node<br>    if (!this.first) {<br>      this.first = node;<br>      this.last = node;<br>    } else {<br>      // Link the current last node to the new node and update last<br>      this.last.next = node;<br>      this.last = node;<br>    }<br><br>    return ++this.size;<br>  }<br><br>  /**<br>   * Removes and returns the element from the front of the queue.<br>   * @returns {*|null} The value of the removed node, or null if queue is empty<br>   */<br>  dequeue() {<br>    if (!this.first) return null;<br><br>    const temp = this.first;<br><br>    // If there&#39;s only one element, clear the last reference<br>    if (this.first == this.last) {<br>      this.last = null;<br>    }<br><br>    // Move first pointer to the next node<br>    this.first = this.first.next;<br>    this.size--;<br><br>    return temp.value;<br>  }<br>}</pre><p>The enqueue method adds a new element to the rear of the queue. It creates a new node and updates the last pointer to reference it. If the queue is empty, both first and last are set to this node. Otherwise, the current last node is linked to the new node, and then last is updated. Since this operation only involves pointer updates and no traversal, it runs in <strong>O(1)</strong> time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3RZjC7VxOGDcvez6NXOa9Q.png" /></figure><p>The dequeue method removes the element from the front of the queue and returns its value. If the queue is empty, it returns null. Otherwise, it updates the first pointer to the next node and decreases the size. If there was only one element in the queue, both first and last are reset to null. This operation also runs in <strong>O(1)</strong> time.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Fkbq2lQIHv1XRHKTWQVdYw.png" /></figure><p>The code for these implementations can be found here: <a href="https://github.com/itstheanurag/medium-blogs-content/tree/main/implementations/ds"><em>code</em></a></p><p>If we were to summarize this implementation, we can say that both stack and queue use nodes connected through references. New elements are always added at the end.</p><p>The real difference lies in how elements are removed.</p><p>In a <strong>stack</strong>, we remove data from the top, so the top pointer moves to the previous node.</p><p>In a <strong>queue</strong>, we remove data from the front, so the first pointer moves forward to the next node</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c74796a5591f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Revisiting Docker after a long time. Everything I read.]]></title>
            <link>https://itstheanurag.medium.com/revisiting-docker-after-a-long-time-everything-i-read-7c8f9a35d84b?source=rss-91a5dc5af290------2</link>
            <guid isPermaLink="false">https://medium.com/p/7c8f9a35d84b</guid>
            <category><![CDATA[docker]]></category>
            <category><![CDATA[containers]]></category>
            <category><![CDATA[docker-compose]]></category>
            <category><![CDATA[docker-networking]]></category>
            <dc:creator><![CDATA[Gaurav Kumar]]></dc:creator>
            <pubDate>Thu, 05 Mar 2026 16:53:08 GMT</pubDate>
            <atom:updated>2026-03-05T16:53:08.666Z</atom:updated>
            <content:encoded><![CDATA[<p>Docker is one of those tools you cannot afford to ignore as a developer.</p><p>I first learned it a few years ago. I understood the basics and could run containers, but I rarely had to set things up from scratch. Most projects at work were already containerized by engineers who had long moved on.</p><p>So Docker became something I “knew,” but didn’t deeply understand.</p><p>Recently, I decided to revisit it properly. Not just the commands, but what’s actually happening under the hood and why Docker became so fundamental in modern development.</p><p>This article is the result of that revisit.</p><h4>Problem</h4><p>In an organization, hundreds of developers work on the same project. Most of these projects were established years ago, often by individuals who are no longer with the organization.</p><p>Even if they’re still around, it’s hard to remember the exact environment that was configured back then. You might be able to find the details, but realistically, when you revisit an old project, you don’t remember the exact libraries, tools, or versions that made it work.</p><p>And software won’t run if the versions don’t match correctly. That’s when the familiar line appears: <strong>“It works on my machine.”</strong> Docker solves this exact problem.</p><p><strong>For this article, I’ll assume that you don’t know anything about Docker.</strong></p><h4>So what the hell is Docker?</h4><p>According to the docs, Docker is an OS level virtualization platform, and to be honest, I didn’t truly understand what that means either. So let’s simplify it.</p><p>At its core, Docker is software that lets you run your application in isolated environments on your machine. Each environment has its own dependencies, libraries, and configuration, separate from everything else on your system.</p><p>It’s like creating a small, lightweight computer inside your computer — but without actually spinning up a full virtual machine.</p><h4>What Is a Virtual Machine? And How Is Docker Different?</h4><p>A virtual machine (VM) is software that emulates an entire computer. Each VM includes its own operating system, its own kernel, and virtualized hardware.</p><p>To make that possible, a component called a <strong>hypervisor</strong> sits between the host system and the virtual machines. The hypervisor manages resources like CPU, memory, and storage, and allows multiple guest operating systems to run on the same physical machine.</p><p>Because every VM runs a full OS, it is heavier and consumes more resources.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VNPlclC94ND0GuFNLku1ZA.png" /><figcaption>VM vs Docker</figcaption></figure><p>Docker takes a different approach.</p><p>Containers do not run a full operating system. There is no guest OS per container and no hypervisor emulating hardware. Instead, Docker Engine runs on the host OS and creates isolated environments that share the host’s kernel.</p><p>That’s why containers are lightweight and start almost instantly.</p><p>In simple terms:</p><p>A virtual machine virtualizes hardware. Docker containers isolate processes at the operating system level.</p><h4>Docker terminology and its core components</h4><p>Now that we have a basic idea about Docker, let’s get acquainted with a few terms that make up the Docker ecosystem.</p><p><strong>Docker Client </strong>is the command-line tool you use to interact with Docker. When you run commands like docker build or docker runThe client sends those instructions to the Docker Engine.</p><p><strong>Docker Engine </strong>is the core runtime that builds images and runs containers. It manages container lifecycle, networking, and storage on your system.</p><p><strong>Dockerfile </strong>is a file that contains instructions for building a Docker image. It defines the base image, dependencies, environment setup, and the command that starts your application.</p><p><strong>Docker Hub </strong>is a public registry where Docker images are stored and shared. You can pull official images like node, nginx, or redis directly from it.</p><p><strong>Docker Registry </strong>is any service that stores Docker images. Docker Hub is the most common public registry, but companies often use private registries to manage internal images securely.</p><h4>Images and containers</h4><p>Before going further, it’s important to clearly understand the difference between an image and a container. These two terms are often used together, but they are not the same thing.</p><p><strong><em>Image</em></strong></p><p>A Docker image is a packaged version of your application.</p><p>It contains your code and everything your app needs to run. The exact versions. The exact setup. Nothing left to chance.</p><p>An image by itself does not run. It’s more like a saved snapshot of your application environment.</p><p>You create it using a Dockerfile, and once it’s built, you can share it with others or store it online so anyone can download and use it.</p><p>In simple terms, an image is a ready-to-use package of your app.</p><p><strong><em>Containers</em></strong></p><p>A container is a running instance of an image.</p><p>When you execute docker runDocker takes the image and creates a container from it. That container becomes a live, isolated process on your machine.</p><p>Unlike images, containers are not read-only. They have a writable layer that allows them to create files, modify data, and maintain state while running.</p><p>You can create multiple containers from the same image, and each container runs independently.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*98DQARVYKkmadxSJKyY0Fw.png" /></figure><h4>Commanding Docker</h4><p>We’ve covered enough theory. Now it’s time to actually use Docker.</p><p>Docker is controlled through the command line. Every time you type a command like docker run or docker build, you’re talking to the Docker Engine and telling it what to do.</p><p>In case you guys don’t have Docker installed, you can install it from <a href="https://docs.docker.com/desktop/setup/install/linux/"><em>here</em></a>.<em> </em>First, make sure Docker is installed:</p><pre>docker --version</pre><p>If Docker is installed correctly, it will print the version number.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/550/1*iN_jxBgFhq24a2WTpuvBVg.png" /><figcaption>docker version result</figcaption></figure><p>Now let’s run your first container:</p><pre>docker run hello-world</pre><p>If the hello-world image does not exist locally, Docker will automatically download it from Docker Hub. Then it will create a container from that image and run it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/918/1*G2HmJNuxguvAg2GVeDmr7Q.png" /><figcaption>docker run hello-world result</figcaption></figure><p>You should see a message confirming that your Docker installation is working properly.</p><p>Now let’s inspect what just happened.</p><p>To see all downloaded images:</p><pre>docker images</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2yvue8F_JhdkaXRy0SDVHw.png" /><figcaption>docker images result</figcaption></figure><p>To see running containers:</p><pre>docker ps</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*f-IoG1UO08IHJYV9VTG0EA.png" /><figcaption>running containers</figcaption></figure><p>To see all containers (including stopped ones):</p><pre>docker ps -a</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*J0Qd5KVwh_ItO9qcaeJUoA.png" /><figcaption>all containers, even stopped ones</figcaption></figure><p>To stop a running container:</p><pre>docker stop &lt;container_id or container name&gt;</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0UriKw7lHeuoq_jHtbUVMQ.png" /><figcaption>stopped a running container</figcaption></figure><p>To remove a container:</p><pre>docker rm &lt;container_id&gt;</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jFpZF9tO6joY6Fxla3dFWA.png" /><figcaption>removed a container</figcaption></figure><p>To remove an image</p><pre>docker rmi &lt;image_id or image_name&gt;</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*8wZ4ushWlLnlcQg_L42pbg.png" /><figcaption>removed images</figcaption></figure><p>It’s not possible to list all the commands here, so I am just gonna add them all with what they do.</p><pre>docker --version               # shows installed Docker version<br>docker version                 # shows client and engine details<br>docker info                    # displays system-wide Docker information<br><br><br>// IMAGES<br><br>docker images                  # lists all local images<br>docker pull &lt;image_name&gt;       # downloads an image from a registry<br>docker build -t name .         # builds an image from a Dockerfile<br>docker rmi &lt;image_id&gt;          # removes an image<br>docker image prune             # removes unused images<br><br>// Container<br><br>docker run &lt;image_name&gt;        # creates and runs a container<br>docker run -d &lt;image_name&gt;     # runs container in detached mode<br>docker run -p 3000:3000 image  # maps ports (host:container)<br>docker run --name myapp image  # assigns a custom container name<br><br>docker ps                      # lists running containers<br>docker ps -a                   # lists all containers (including stopped)<br>docker stop &lt;container_id&gt;     # stops a running container<br>docker start &lt;container_id&gt;    # starts a stopped container<br>docker restart &lt;container_id&gt;  # restarts a container<br>docker rm &lt;container_id&gt;       # removes a container<br>docker container prune         # removes all stopped containers<br><br>// Inside Container<br><br>docker exec -it &lt;container&gt; bash  # opens interactive shell inside container<br>docker logs &lt;container&gt;           # shows container logs<br>docker logs -f &lt;container&gt;        # follows logs in real time<br>docker inspect &lt;container&gt;        # shows detailed container info<br><br>// Volumes<br><br>docker volume ls               # lists volumes<br>docker volume create name      # creates a volume<br>docker volume rm name          # removes a volume<br><br>// Networks<br><br>docker network ls              # lists networks<br>docker network create name     # creates a network<br>docker network rm name         # removes a network<br><br>// Clean up<br>docker system df               # shows disk usage<br>docker system prune            # removes unused data (containers, networks, images)<br>docker system prune -a         # removes all unused images as well</pre><p>Here’s the official cheat sheet for the Docker commands <a href="https://docs.docker.com/get-started/docker_cheatsheet.pdf"><em>cheatsheet</em></a></p><h4>What is Port Mapping?</h4><p>So you ran your first container:</p><pre>docker run n8nio/n8n:latest</pre><p>You see logs in the terminal:</p><pre>n8n ready on ::, port 5678<br>Editor is now accessible via:<br>http://localhost:5678</pre><p>Naturally, you open your browser and go to <a href="http://localhost:5678.">http://localhost:5678.</a></p><p>And… nothing loads.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FgPhpREtCvDwugT978FhTw.png" /></figure><p>Why doesn’t it work?</p><p>Because the application isn’t running directly on your machine. It’s running inside the container.</p><p>When n8n says it’s listening on port 5678, it means port 5678 inside the container, not on your laptop. From your system’s perspective, nothing is listening on port 5678 at all.</p><p>Containers are isolated by design. That isolation includes networking. They do not automatically expose their internal ports to the host machine.</p><p>So even though the logs confidently say “Editor is now accessible via http://localhost:5678”, your browser cannot reach it.</p><p>This is where port mapping comes in.</p><p>Port mapping connects a port on your machine to a port inside the container. In simple terms, you are telling Docker to forward traffic from your host to the container.</p><p>Once that connection is made, your browser finally knows where to send the request.</p><pre>docker run -p &lt;host_port&gt;:&lt;container_port&gt; &lt;image_name&gt; // structure<br><br>docker run -p 5678:5678 n8nio/n8n:latest // actual command</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DO4Mk2UjbrtPcpcyKb2E6Q.png" /></figure><p>Now it works as it should.</p><h4>Persisting the Data (Volumes)</h4><p>At this point, your container is running and accessible. You mapped the port, opened the app in your browser, and everything works as expected.</p><p>You create workflows in n8n. You configure settings. You test things. It feels solid.</p><p>Then you stop the container.</p><pre>docker stop &lt;container_name&gt;</pre><p>If you start the container again, the data is lost.</p><p><strong><em>Why does this happen?</em></strong></p><p>Because containers are temporary by nature.</p><p>When you run a container, any data created inside it stays inside that container. If you delete the container, everything stored inside it is deleted as well.</p><p>The image still exists. But the container — and its data — are gone.</p><p>This behavior is intentional. Containers are designed to be easy to create, remove, and recreate. That works well for applications that don’t store data.</p><p>But you need a way to store the data, and Docker provides a few for us.</p><p><strong><em>Container Writable Layer (Temporary)</em></strong></p><p>This is the default.</p><p>Data lives inside the container and disappears when the container is removed.</p><p>Good for:</p><ul><li>Temporary testing</li><li>Stateless apps</li></ul><p>Not good for:</p><ul><li>Databases</li><li>Uploaded files</li><li>Configuration storage</li></ul><p><strong><em>Bind Mounts</em></strong></p><p>A bind mount connects a specific folder on your host machine to a folder inside the container.</p><p>Example:</p><pre>docker run -v /home/user/data:/app/data my-image</pre><p>Here:</p><ul><li>/home/user/data → folder on your machine</li><li>/app/data → folder inside the container</li></ul><p>The container writes directly to your local filesystem.</p><p>Good for:</p><ul><li>Development</li><li>Local file syncing</li></ul><p>Downside:</p><ul><li>Tightly coupled to the host filesystem</li><li>Less portable</li></ul><p><strong><em>Volumes (Recommended for Most Cases)</em></strong></p><p>A volume is a storage managed by Docker.</p><p>It exists independently of containers and is stored in Docker’s internal storage area.</p><p>Example:</p><pre>docker volume create mydata<br>docker run -v mydata:/app/data my-image</pre><p>The container can read and write to it, but deleting the container does not delete the volume.</p><p>Good for:</p><ul><li>Databases</li><li>Production setups</li><li>Clean separation of data and containers</li></ul><h4>Networking In Docker</h4><p>Docker containers are isolated environments that run independently of each other. That isolation is intentional.</p><p>But real-world systems are rarely a single container. A backend application might depend on a database, a cache like Redis, a message queue, and maybe a monitoring service.</p><p>All of these services need to communicate with each other.</p><p>Docker makes this possible through networks.</p><p>By default, Docker attaches containers to a network called bridge. Containers on the same network can communicate with each other using their container names as hostnames.</p><p>But in real-world setups, relying on the default network is not ideal. You usually want to define your own custom network.</p><p>For example, suppose you’re building a fitness app backend. You might run:</p><ul><li>A Node.js API</li><li>A PostgreSQL database</li><li>A Redis instance</li><li>A background worker</li></ul><p>Instead of letting all of them sit on the default bridge network, you can create your own isolated network:</p><pre>docker network create fitness-network</pre><p>Then attach each container to it:</p><pre>docker run -d --name api --network fitness-network my-api<br>docker run -d --name db --network fitness-network postgres<br>docker run -d --name redis --network fitness-network redis</pre><p>Now all these containers can talk to each other securely within fitness-network.</p><p><strong><em>Docker Network Drivers</em></strong></p><p>When you create a network, Docker uses something called a <strong>network driver</strong>. The driver determines how containers communicate.</p><p>Some common drivers are:</p><ul><li><strong>bridge</strong> — The default driver. Used for communication between containers on a single host.</li><li><strong>host</strong> — Removes network isolation and lets the container use the host’s network directly.</li><li><strong>none</strong> — Disables networking completely.</li><li><strong>overlay</strong> — Used for communication between containers running on different Docker hosts (commonly in swarm mode).</li><li><strong>macvlan — </strong>Assigns a MAC address to containers, making them appear as physical devices on the network.</li><li><strong>ipvlan — </strong>Similar to macvlan, but uses IP-based routing for advanced networking setups.</li></ul><p>When you create a custom network without specifying a driver, Docker uses the bridge driver by default.</p><p>For example:</p><pre>docker network create --driver bridge fitness-network</pre><h4>Containerizing an Application ( Dockerfile )</h4><p>Theory is good, but Docker becomes real when you containerize your own application.</p><p>We do this by writing a Dockerfile in the root of the project. A Dockerfile is simply a set of instructions that tells Docker how to build an image:</p><ul><li>Which base image to use</li><li>What dependencies to install</li><li>What files to copy</li><li>What command should start the application</li></ul><p>It’s a recipe for your application’s environment. For this article, I created a simple Hono + bun-based backend project that exposes only two endpoints, and I built an image from it. <a href="https://github.com/itstheanurag/medium-blogs-content/tree/main/docker"><em>link</em></a></p><pre>FROM oven/bun:latest</pre><pre>WORKDIR /app</pre><pre>COPY package.json bun.lock ./<br>RUN bun install</pre><pre>COPY . .<br></pre><pre>EXPOSE 3000</pre><pre>CMD [&quot;bun&quot;, &quot;start&quot;]</pre><pre>// hello-hono:latest - name  489MB - image size    132MB - content size</pre><p>This is a very basic Dockerfile that builds an image from our application..</p><h4>Optimizing Images</h4><p>When we talk about optimizing a Docker image, we’re not just trying to make it smaller. We’re trying to make it lean, secure, and efficient.</p><p>The first and most impactful decision is choosing the right base image.</p><p>Take Node.js as an example. There isn’t just one Node image. There are multiple variants — some large, some minimal.</p><p>You’ll commonly see options like:</p><ul><li>node:20</li><li>node:20-slim</li><li>node:20-alpine</li></ul><p>They all run Node. But they are not the same.</p><p>The standard node Image is Debian-based and includes many system utilities. It’s convenient and predictable, but larger in size.</p><p>The slim variant removes unnecessary packages while keeping the essential runtime components.</p><p>The alpine variant is built on Alpine Linux, which is extremely small. This can drastically reduce your final image size.</p><p>Choosing the right base image alone can reduce your image by hundreds of megabytes. But it&#39;s only the first step.</p><p>You need to make sure that your project image stays as lean as possible. To make sure that happens, we need to understand layers in Docker</p><p><strong><em>Understanding Docker Layers</em></strong></p><p>Docker images are built in layers. Each instruction in a Dockerfile creates a new layer, and these layers are cached.</p><p>For example:</p><pre>FROM → base layer  <br>WORKDIR → new layer  <br>COPY package.json bun.lock → new layer  <br>RUN bun install → new layer  <br>COPY . . → new layer  </pre><p>Each of these steps becomes part of the final image.</p><p>When you rebuild an image, Docker reuses the cached layers until it reaches a step where something has changed. From that point onward, it rebuilds only that layer and everything that follows. Everything before that is reused.</p><p>This caching mechanism is powerful — but only if you structure your Dockerfile intentionally.</p><p><strong><em>Structure Instructions for Caching</em></strong></p><p>Bad pattern:</p><pre>COPY . .<br>RUN bun install</pre><p>Every time you change a single source file, dependencies are reinstalled.</p><p>Better pattern:</p><pre>COPY package.json bun.lock ./<br>RUN bun install<br>COPY . .</pre><p>Now dependencies are only reinstalled when your lock file changes.</p><p><strong><em>Reduce Layer Bloat</em></strong></p><p>Each instruction creates a new layer. If temporary files or package caches are created in one layer and cleaned in another, they remain in the image history.</p><p>For example:</p><pre>RUN apt update<br>RUN apt install -y curl<br>RUN apt clean</pre><p>Three layers.</p><p>Better:</p><pre>RUN apt update &amp;&amp; \<br>    apt install -y curl &amp;&amp; \<br>    apt clean &amp;&amp; \<br>    rm -rf /var/lib/apt/lists/*</pre><p>Now everything happens in one layer, and temporary data is removed before the layer is finalized.</p><p><strong><em>Control the Build Context with </em></strong><strong><em>.dockerignore</em></strong></p><p>Even before layers are created, your entire project directory is sent to Docker as the build context.</p><p>.dockerignore allows us to control what is sent in the Docker context for the build process. If you don’t define .dockerignore, you might accidentally include:</p><ul><li>.git</li><li>local node_modules</li><li>logs</li><li>.env files</li></ul><p>That increases build time and risks leaking secrets.</p><p>A simple .dockerignore:</p><pre>node_modules<br>.git<br>.env</pre><p><strong><em>Separate Build and Runtime (Multi-Stage Builds)</em></strong></p><p>The most powerful optimization technique is separating build-time dependencies from runtime.</p><p>Your production container should not contain:</p><ul><li>TypeScript compiler</li><li>Dev dependencies</li><li>Testing libraries</li><li>Build tools</li></ul><p>Only the compiled output and runtime.</p><p>This reduces:</p><ul><li>Image size</li><li>Attack surface</li><li>Deployment time</li></ul><h4>Run with Least Privilege</h4><p>By default, most Docker containers run as the root user inside the container.</p><p>That sounds harmless at first. After all, it’s “just a container,” right?</p><p>But remember — containers share the host kernel. If your application is compromised and it’s running as root, the attacker now has elevated privileges inside that container. That’s not something you want to gamble with.</p><p>So instead of relying on the default behavior, we explicitly create a non-root user for the application.</p><p>Below is an example from a side project ( which I never shipped )</p><pre>RUN addgroup --system --gid 1001 nodejs<br>RUN adduser --system --uid 1001 nextjs</pre><p>What’s happening here?</p><p>We’re creating:</p><ul><li>A system group called nodejs</li><li>A system user called nextjs</li></ul><p>This user exists only to run the application. Nothing more.</p><p>But here’s where people usually stop — and that’s a mistake.</p><p>Creating a non-root user is not enough. If you switch to that user without adjusting file ownership, your app will fail with permission errors. Why? Because files copied into the container are owned by root default.</p><p>That’s why we assign ownership explicitly while copying build artifacts:</p><pre>COPY --from=builder /app/public ./public<br>COPY --from=builder --chown=nextjs:nodejs /app/.next/standalone ./<br>COPY --from=builder --chown=nextjs:nodejs /app/.next/static ./.next/static</pre><p>Now the important runtime files belong to nextjs:nodejs.</p><p>The application can read and execute what it needs. It doesn’t need root. It doesn’t get root.</p><p>Finally, we switch the runtime user:</p><pre>USER nextjs</pre><p>At this point, the container runs with limited privileges.</p><h4>Composing An Entire Application</h4><p>Now that we have a decent idea about images and how to build them, we have also built a decent small image and run it in a container. But real applications are rarely just one container.</p><p>Let’s take a simple backend system as an example.</p><p>Your API might need:</p><ul><li>A PostgreSQL database</li><li>A metrics system like Prometheus</li><li>A visualization layer like Grafana</li><li>A log aggregator like Loki</li></ul><p>Now imagine managing all of this manually.</p><p>You would have to:</p><ol><li>Create a network</li><li>Start PostgreSQL with environment variables</li><li>Attach it to the network</li><li>Create a volume for persistent storage</li><li>Start Prometheus with its config file</li><li>Start Grafana</li><li>Link Grafana to Prometheus</li><li>Then, finally, start your backend</li></ol><p>That’s not just inconvenient. It’s fragile.</p><p>If someone else clones your project, they must replicate every one of those steps exactly.</p><p>This is the problem Docker Compose solves.</p><p><strong><em>Docker Compose</em></strong></p><p>Docker Compose is a tool that allows you to define and run multi-container applications using a single configuration file — usually named docker-compose.yml.</p><p>You basically describe everything that is required to run your application in a single yaml file, instead of manually starting all the containers one by one through cli.</p><p>You define:</p><ul><li>What services exist</li><li>Which images do they use</li><li>Which ports do they expose</li><li>Which volumes persist data</li><li>Which networks do they join</li><li>How services depend on each other</li></ul><p>Then Docker Compose orchestrates all of it for you. Here’s an example ( <a href="https://github.com/ravixstudio/backend-template/blob/main/docker-compose.dev.yml"><em>full file</em></a> )</p><pre>services:<br>  postgres:<br>    image: &quot;postgres&quot;<br>    container_name: &quot;backend-template-postgres&quot;<br>    environment:<br>      POSTGRES_USER: &quot;postgres&quot;<br>      POSTGRES_PASSWORD: &quot;postgres&quot;<br>      POSTGRES_DB: &quot;backend-template&quot;<br>    ports:<br>      - &quot;${POSTGRES_PORT:-5432}:5432&quot;<br>    volumes:<br>      - postgres-data:/var/lib/postgresql<br>    restart: always<br>    networks:<br>      - backend-network</pre><p>Let’s break this down.</p><ul><li>image: Pulls the official PostgreSQL image.</li><li>environment: Configures database credentials and name.</li><li>ports: Maps port 5432 inside the container to the host.</li><li>volumes: Ensures database data persists even if the container is removed.</li><li>networks: Connects the service to a shared internal network.</li></ul><p>This single block replaces a long docker run command with multiple flags.</p><p>Here are a few Docker Compose commands:</p><p>Start everything:</p><pre>docker compose up</pre><p>Run in detached mode:</p><pre>docker compose up -d</pre><p>Stop everything:</p><pre>docker compose down</pre><p>Rebuild images if needed:</p><pre>docker compose up --build</pre><p>Remove containers, networks, and volumes:</p><pre>docker compose down -v</pre><h4>conclusion</h4><p>I think this article is in a good place to end. Before I leave, here’s one final thought.</p><p>Thinking you’ve mastered something just because you learned it years ago — and have been using it unconsciously ever since — can be misleading.</p><p>Familiarity is not the same as understanding.</p><p>It’s only when you go back, strip it down to first principles, and rebuild it carefully that you realize how much you were operating on autopilot.</p><p>Revisiting Docker reminded me of that. And that lesson probably applies to more than just Docker ( JavaScript, since I have been trying to create videos on it on YouTube for absolute freshers, and it’s damn hard )</p><p>Follow me on X if you like: <a href="https://x.com/Itstheanurag"><strong><em>Profile</em></strong></a></p><p>More from me:</p><ul><li><a href="https://medium.com/@itstheanurag/redis-all-that-you-need-to-know-as-a-developer-3577a29e4698"><em>All you need to know about Redis</em></a></li><li><a href="https://medium.com/@itstheanurag/one-line-millions-of-operations-the-world-behind-print-hello-world-749cf42df292">under the hood of print(“Hello world”)</a></li><li><a href="https://medium.com/@itstheanurag/understanding-caching-more-than-just-redis-3bb746628337">Understanding caching — more than Redis</a></li><li><a href="https://medium.com/@itstheanurag/authentication-in-a-production-grade-system-design-implementation-9aef90fb0dde">Authentication in a Production Grade system</a></li><li><a href="https://medium.com/@itstheanurag/what-are-streams-and-how-to-use-them-to-build-reliable-backend-systems-7a1b3e583305">What are streams, and how to use them to build reliable backend systems.</a></li><li><a href="https://medium.com/@itstheanurag/understanding-polling-long-polling-sse-and-websockets-when-to-use-what-63bf294460f2">Understanding Polling, Long Polling, SSE, and WebSockets: When to Use What</a></li><li><a href="https://medium.com/@itstheanurag/code-alone-isnt-enough-for-production-systems-5fe7e485bd96">Writing code alone isn’t enough for building production grade system</a></li><li><a href="https://medium.com/@itstheanurag/the-problems-i-am-facing-while-building-a-shadcn-like-system-for-the-backend-33aff382b241">Building Hanma, the Shadcn for backend coding.</a></li><li><a href="https://medium.com/@itstheanurag/stop-uploading-files-to-the-server-unnecessarily-c1a0152b3efd">Stop uploading files to the server unnecessaril</a>y</li><li><a href="https://medium.com/@itstheanurag/backend-development-more-than-endpoints-43a330b6e006">Backend development: more than just CRUDs</a></li><li><a href="https://medium.com/@itstheanurag/a-walk-through-arrays-in-javascript-bd994e9d5422">A walk through arrays in JavaScrip</a>t</li><li><a href="https://medium.com/@itstheanurag/building-a-shadcn-for-the-backend-an-overview-of-hanma-567fa5ef0302">Building Shadcn for the backend, an overview of Hanma</a></li><li><a href="https://medium.com/@itstheanurag/the-problems-i-am-facing-while-building-a-shadcn-like-system-for-the-backend-33aff382b241">The problems I am facing while building Hanma</a></li><li><a href="https://medium.com/@itstheanurag/understanding-the-conversions-in-javascript-10fa453f2982">Understanding the conversions in JavaScript.</a></li><li><a href="https://medium.com/@itstheanurag/functions-are-crazy-in-javascript-4978d4d6105b">Functions are crazy in JavaScript</a></li><li><a href="https://medium.com/@itstheanurag/backend-development-more-than-endpoints-43a330b6e006">Backend development: More than endpoints</a></li><li><a href="https://medium.com/@itstheanurag/a-deep-dive-into-console-in-javascript-e93212d44e9f">A deep dive into JavaScript Console</a></li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7c8f9a35d84b" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>