<?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 Aditi Dosi on Medium]]></title>
        <description><![CDATA[Stories by Aditi Dosi on Medium]]></description>
        <link>https://medium.com/@adosi12?source=rss-7b54740e75bb------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*W6o4FYBAoG0HQKyKhk1Nmg.png</url>
            <title>Stories by Aditi Dosi on Medium</title>
            <link>https://medium.com/@adosi12?source=rss-7b54740e75bb------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 14:08:32 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@adosi12/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[Reinventing Myself: From Burnout to Growth]]></title>
            <link>https://adosi12.medium.com/reinventing-myself-from-burnout-to-growth-fe1b68c18ea2?source=rss-7b54740e75bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/fe1b68c18ea2</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[life]]></category>
            <category><![CDATA[restart]]></category>
            <category><![CDATA[career-advice]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Aditi Dosi]]></dc:creator>
            <pubDate>Sun, 30 Aug 2026 06:49:34 GMT</pubDate>
            <atom:updated>2026-08-30T06:49:34.240Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*2nZ7Nvp7Irt01PdpjIPp-g.png" /></figure><h3>I Have a Job, But I Don’t Feel Like I’m Living My Life</h3><p><em>Three years into my corporate career, I realized that being stable and being fulfilled are two very different things.</em></p><p>Three years ago, getting a job felt like an achievement.</p><p>I had entered the corporate world, started earning, gained experience, and slowly built a routine around my work. From the outside, everything looked fine.</p><p>But lately, I have been feeling something I cannot ignore anymore.</p><p>I feel anxious almost every day.</p><p>Sometimes, I suddenly feel like crying.</p><p>And the hardest part is that there isn’t always one specific reason.</p><p>It feels like something inside me is constantly telling me:</p><p><strong>This is not where you want to stay forever.</strong></p><h3>When Your Job Becomes a Routine Instead of a Dream</h3><p>I don’t hate working.</p><p>I don’t hate learning.</p><p>But somewhere along the way, I stopped enjoying what I do.</p><p>Every morning started feeling similar. Work, deadlines, meetings, tasks, sleep — and then the same cycle again.</p><p>And while I continued doing my job, another part of me started becoming louder.</p><p>A part of me wanted to learn again.</p><p>I wanted to explore Artificial Intelligence.</p><p>I wanted to restart Data Structures and Algorithms.</p><p>I wanted to build things.</p><p>I wanted to become better — not because someone else expected it from me, but because I knew I was capable of more.</p><p>The problem was not that I didn’t know what I wanted.</p><p>The problem was that I couldn’t stay consistent.</p><h3>The Constant Battle Between Work, Learning, and Guilt</h3><p>Some days, office work takes all my energy.</p><p>By the time I finish, I don’t feel like opening another laptop screen to study.</p><p>And then there are days when I actually have time, but I still don’t study.</p><p>Maybe I’m tired.</p><p>Maybe I’m overwhelmed.</p><p>Maybe I’m just mentally exhausted.</p><p>But instead of accepting that, I feel guilty.</p><p>Then the overthinking starts.</p><p><strong>“Everyone else is moving ahead.”</strong></p><p><strong>“People are learning AI, building projects, solving DSA problems, and switching to better opportunities.”</strong></p><p><strong>“What am I doing?”</strong></p><p>And suddenly, one day of not studying starts feeling like I’m destroying my future.</p><p>That’s when anxiety becomes the loudest.</p><p>I start thinking about everything I want to achieve this year.</p><p>I think about the skills I need to learn.</p><p>The projects I haven’t started.</p><p>The DSA problems I haven’t solved.</p><p>The AI concepts I don’t understand yet.</p><p>And instead of motivating me, the long list freezes me.</p><h3>Maybe I’m Not Lazy. Maybe I’m Just Overwhelmed.</h3><p>For a long time, I thought my biggest problem was laziness.</p><p>But now I think it might be something else.</p><p>Maybe I am overwhelmed by the pressure I put on myself.</p><p>I want to improve my career.</p><p>I want to learn AI.</p><p>I want to restart DSA.</p><p>I want to become more disciplined.</p><p>I want to create a better future.</p><p>And somehow, I expect myself to do everything perfectly while also working a full-time job and handling life.</p><p>No wonder I feel exhausted.</p><p>No wonder I sometimes want to stop everything and cry.</p><p>We often talk about productivity, discipline, and consistency.</p><p>But we don’t talk enough about how difficult it can be to chase a new version of yourself while you’re still trying to survive your current life.</p><h3>The Fear of Being Left Behind</h3><p>I think one of my biggest fears is being left behind.</p><p>Technology is moving fast.</p><p>AI is changing industries.</p><p>Every day, there is a new tool, model, framework, or skill to learn.</p><p>And when you’re already struggling to find time and energy, the speed of everything can feel terrifying.</p><p>You start asking yourself:</p><p><strong>Will I be able to catch up?</strong></p><p><strong>Will I achieve what I want this year?</strong></p><p><strong>What if I waste another year doing nothing?</strong></p><p>But maybe these questions are part of the problem.</p><p>Maybe I don’t need to catch up with everyone.</p><p>Maybe I just need to move forward from where I am.</p><p>Because the truth is, I am not starting from zero.</p><p>Three years of working have taught me discipline.</p><p>They have taught me how companies work.</p><p>They have taught me responsibility, pressure, communication, and problem-solving.</p><p>Even the years that now feel confusing were not wasted.</p><p>They brought me here.</p><p>And now, I know something about myself that I didn’t know before:</p><p><strong>I want more from my career.</strong></p><p>And recognizing that is the beginning of change.</p><h3>I Am Learning to Start Again Without Hating Myself</h3><p>Maybe I won’t study for five hours every day.</p><p>Maybe I will miss a few days.</p><p>Maybe I will restart DSA and forget concepts again.</p><p>Maybe I will explore AI slowly instead of becoming an expert overnight.</p><p>And that’s okay.</p><p>I’m starting to understand that consistency doesn’t mean being perfect.</p><p>Consistency means returning.</p><p>Returning after a busy week.</p><p>Returning after losing motivation.</p><p>Returning after taking a break.</p><p>Returning even when you feel like you’ve wasted time.</p><p>One DSA problem is still progress.</p><p>One hour of learning about AI is still progress.</p><p>One small project is still progress.</p><p>The problem is that we often don’t value small progress because we’re too focused on the final destination.</p><p>We want the better job.</p><p>The better salary.</p><p>The new skills.</p><p>The confidence.</p><p>The successful version of ourselves.</p><p>But we forget that person is built on ordinary days.</p><p>On tired evenings.</p><p>On small efforts.</p><p>On restarting again and again.</p><h3>Maybe This Is What Growth Actually Looks Like</h3><p>Right now, I don’t have everything figured out.</p><p>I still feel anxious sometimes.</p><p>I still have days when I don’t study.</p><p>I still worry about whether I will achieve everything I want in the coming years.</p><p>But I don’t want anxiety to convince me that I am failing.</p><p>I don’t want one unproductive day to make me believe that I have no future.</p><p>And I don’t want to spend my entire life waiting until I feel completely ready.</p><p>So, I’m choosing something simpler.</p><p>I will learn.</p><p>Slowly.</p><p>I will explore AI with curiosity instead of fear.</p><p>I will restart DSA without being ashamed of what I have forgotten.</p><p>I will take breaks when I need them.</p><p>And when I lose momentum, I will start again.</p><p>Because maybe becoming a better version of yourself doesn’t mean transforming your life in one year.</p><p>Maybe it simply means refusing to give up on yourself.</p><h3>I’m Not Behind. I’m Just Beginning Again.</h3><p>I may be three years into my career.</p><p>I may feel confused.</p><p>I may not know exactly where I’ll be next year.</p><p>But I know that the desire to grow is still inside me.</p><p>And as long as that desire exists, it’s never too late to begin again.</p><p>So this is where I am right now.</p><p>A little tired.</p><p>A little anxious.</p><p>A little lost.</p><p>But still hopeful.</p><p>Still curious.</p><p>Still willing to learn.</p><p>And maybe that’s enough.</p><p><strong>Because restarting doesn’t mean you failed.</strong></p><p><strong>Sometimes, restarting means you finally decided not to settle.</strong></p><p>#CareerGrowth #CareerChange #CorporateLife #WorkAnxiety #CareerJourney #FeelingLost #SelfGrowth #PersonalGrowth #LearningJourney #ArtificialIntelligence #AIJourney #DSA #SoftwareEngineering #TechCareer #DeveloperLife #Consistency #GrowthMindset #StartAgain #NeverTooLate #BetterVersionOfMe #MentalHealthMatters #LifeInYourTwenties</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fe1b68c18ea2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[10 Techy Hacks in System Design That Power Real-World Scale ]]></title>
            <link>https://adosi12.medium.com/10-techy-hacks-in-system-design-that-power-real-world-scale-008c0338b169?source=rss-7b54740e75bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/008c0338b169</guid>
            <category><![CDATA[system-design-interview]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[women-in-tech]]></category>
            <category><![CDATA[scalability]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Aditi Dosi]]></dc:creator>
            <pubDate>Thu, 27 Aug 2026 07:17:22 GMT</pubDate>
            <atom:updated>2026-08-27T07:17:22.540Z</atom:updated>
            <content:encoded><![CDATA[<p><em>From the 10:00 AM Tatkal rush to instant payments, food delivery, streaming, and global SaaS — these are the system-design tricks hiding behind everyday products.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-pQVVQtC7FI6wLFKK6Cudg.png" /></figure><p>System design is often taught using boxes:</p><p><strong>Client → Load Balancer → API → Database.</strong></p><p>It looks simple.</p><p>Until <strong>millions of users click the same button at the same time.</strong></p><p>A production system has to deal with traffic spikes, retries, slow databases, network failures, stale data, unpredictable workloads, and users who absolutely refuse to wait.</p><p>That is where interesting system-design decisions begin.</p><p>Instead of looking at system design as a collection of theoretical components, let’s look at <strong>10 practical hacks inspired by real-world scenarios</strong>.</p><blockquote><strong><em>The goal isn’t to copy these architectures.</em></strong></blockquote><blockquote><em>The goal is to recognize the underlying problem and know which tool to reach for.</em></blockquote><h3>1. The 10:00 AM Problem: Don’t Let Everyone Enter at Once</h3><h3>The scenario: IRCTC Tatkal</h3><p>Imagine a ticket booking system where a huge number of users know that bookings open at a particular time.</p><p>For Tatkal, AC-class bookings open at <strong>10:00 AM</strong>, while non-AC Tatkal bookings open at <strong>11:00 AM</strong>.</p><p>Now imagine thousands or millions of clients effectively saying:</p><blockquote><strong><em>“BOOK NOW!”</em></strong></blockquote><p>at almost exactly the same moment.</p><p>If every request immediately hits the application servers and database, you don’t have a traffic problem anymore.</p><p>You have a <strong>synchronization problem</strong>.</p><h3>The hack: Queue the traffic</h3><p>Instead of:</p><pre>Millions of Users<br>       ↓<br>    API Server<br>       ↓<br>    Database</pre><p>think:</p><pre>Millions of Users<br>       ↓<br>   API Gateway<br>       ↓<br> Rate Limiter<br>       ↓<br> Virtual Queue<br>       ↓<br>    Workers<br>       ↓<br> Booking Service<br>       ↓<br>    Database</pre><p>The queue acts as a <strong>shock absorber</strong>.</p><p>Users can arrive at an enormous rate, while the backend processes requests at a rate it can actually sustain.</p><h3>Add a token bucket</h3><p>A <strong>token bucket</strong> can control how quickly requests are released.</p><p>For example:</p><pre>Request<br>             ↓<br>      ┌──────────────┐<br>      │ Token Bucket │<br>      └──────┬───────┘<br>             ↓<br>      Allowed Requests<br>             ↓<br>          Workers</pre><p>Tokens are generated at a controlled rate.</p><p>If tokens exist → process the request.</p><p>If not → wait or reject.</p><h3>Why this works</h3><p>The key insight is:</p><p><strong>Don’t make your database absorb your traffic spike.</strong></p><p>Make another layer absorb it.</p><p>This pattern applies to:</p><ul><li>Ticket booking</li><li>Concert ticket sales</li><li>Flash sales</li><li>Limited-edition product launches</li><li>Government portals</li><li>Exam registrations</li></ul><p>And importantly, <strong>this is a system-design lesson inspired by the Tatkal traffic pattern — not a claim about IRCTC’s exact private architecture.</strong></p><h3>2. The 11:00 AM Problem: Traffic Doesn’t Always Disappear After the Spike</h3><p>Here’s a subtle problem.</p><p>Suppose the huge spike happens at 10:00 AM.</p><p>You design your infrastructure for that spike.</p><p>Great.</p><p>But what happens at <strong>10:05, 10:30, or 11:00?</strong></p><p>Traffic patterns can change again.</p><p>This is where <strong>autoscaling + predictive scaling</strong> becomes useful.</p><h3>The hack: Scale before users arrive</h3><p>Traditional autoscaling reacts to metrics:</p><pre>Traffic ↑<br>   ↓<br>CPU ↑<br>   ↓<br>Autoscaler notices<br>   ↓<br>New servers</pre><p>That’s reactive.</p><p>But predictable events allow something smarter:</p><pre>Historical Data<br>      +<br>Calendar/Event Data<br>      ↓<br>Traffic Prediction<br>      ↓<br>Pre-scale Infrastructure<br>      ↓<br>Traffic Arrives</pre><p>If you know a sale begins at 10:00 AM, why wait until CPU reaches 90%?</p><p><strong>Prepare before the traffic arrives.</strong></p><h3>This is useful for</h3><ul><li>E-commerce sales</li><li>Ticketing</li><li>Tax portals</li><li>Cricket match traffic</li><li>Food delivery during lunch</li><li>OTT platforms during major releases</li></ul><p>The trick is simple:</p><blockquote><strong><em>Use historical traffic as a signal, not just current CPU.</em></strong></blockquote><h3>3. Cache What Doesn’t Need to Be Calculated Again</h3><p>Let’s say an e-commerce application receives:</p><p><strong>10 million product-page requests.</strong></p><p>Does every request need to hit the database?</p><p>Probably not.</p><p>Product name?</p><p>Probably unchanged.</p><p>Product image?</p><p>Definitely cacheable.</p><p>Category?</p><p>Probably unchanged.</p><p>Configuration?</p><p>Often cacheable.</p><h3>The hack: Cache-aside</h3><p>Instead of:</p><pre>User → Application → Database</pre><p>use:</p><pre>User<br> ↓<br>Application<br> ↓<br>Cache ─── HIT ──→ Return<br> ↓<br>MISS<br> ↓<br>Database<br> ↓<br>Cache<br> ↓<br>Return</pre><p>A popular implementation is the <strong>cache-aside pattern</strong>.</p><p>The application first checks the cache.</p><p>If the value exists:</p><p><strong>Cache Hit → return immediately.</strong></p><p>If it doesn’t:</p><p><strong>Cache Miss → query DB → populate cache.</strong></p><h3>But caching introduces a new problem</h3><p><strong>Stale data.</strong></p><p>Suppose a product price changes.</p><p>Your cache still says:</p><blockquote><em>₹999</em></blockquote><p>while the database says:</p><blockquote><em>₹899</em></blockquote><p>Now you need a strategy:</p><ul><li>TTL</li><li>Explicit invalidation</li><li>Versioned keys</li><li>Stale-while-revalidate</li><li>Event-based invalidation</li></ul><p>The real system-design lesson is:</p><blockquote><strong><em>Caching is easy. Cache invalidation is architecture.</em></strong></blockquote><h3>4. Don’t Make Users Wait for Work They Don’t Care About</h3><p>Consider ordering food. You place an order.</p><p>Several things may need to happen:</p><pre>Order<br> ├── Payment<br> ├── Restaurant notification<br> ├── Delivery assignment<br> ├── Email<br> ├── SMS<br> └── Analytics</pre><p>Should the user wait for every operation? No.</p><h3>The hack: Event-driven architecture</h3><p>The critical path might be:</p><pre>User<br> ↓<br>Order Service<br> ↓<br>Payment<br> ↓<br>Order Confirmed</pre><p>Then publish an event:</p><pre>OrderCreated<br>     ↓<br>   Kafka<br>  ↙  ↓  ↘<br>Email  Analytics  Notification</pre><p>Now those services can work independently.</p><h3>Why this is powerful</h3><p>Without a queue:</p><pre>A → B → C → D → E</pre><p>If D is slow, everything waits.</p><p>With asynchronous processing:</p><pre>→ B<br>A → Queue → C<br>        → D<br>        → E</pre><p>Services become <strong>loosely coupled</strong>.</p><p>This pattern appears everywhere:</p><ul><li>Order processing</li><li>Notifications</li><li>Payment events</li><li>Analytics</li><li>Audit logs</li><li>Image processing</li><li>Video transcoding</li></ul><p>The trick:</p><blockquote><strong><em>If the user doesn’t need the result immediately, don’t put it on the critical path.</em></strong></blockquote><h3>5. Payments Have a Special Problem: “Retry”</h3><p>Imagine you pay ₹2,000.</p><p>Your bank processes the payment.</p><p>But your app times out.</p><p>You don’t know whether it succeeded.</p><p>So you press: <strong>Pay Again.</strong></p><p>Now the server receives the same operation twice. Without protection:</p><pre>Payment Request<br>      ↓<br>Charge ₹2,000</pre><pre>Retry<br>      ↓<br>Charge ₹2,000 AGAIN 😱</pre><h3>The hack: Idempotency keys</h3><p>The client generates a unique request ID:</p><pre>idempotency_key = abc123</pre><p>The server stores the result:</p><pre>abc123 → SUCCESS</pre><p>If the same request arrives again:</p><pre>abc123 already processed<br>        ↓<br>Return previous result</pre><p>Instead of executing the payment again.</p><h3>Conceptually</h3><pre>Client<br>  ↓<br>Request + Idempotency Key<br>  ↓<br>Idempotency Store<br>  ↓<br>Already processed?<br>   ↙       ↘<br> YES       NO<br>  ↓         ↓<br>Return    Process<br>result      ↓<br>          Store<br>          result</pre><p>This is one of the most important patterns in distributed systems.</p><p>Because:</p><blockquote><strong><em>Retries are inevitable. Duplicate side effects are not.</em></strong></blockquote><p>Use it for:</p><ul><li>Payments</li><li>Ticket booking</li><li>Orders</li><li>Fund transfers</li><li>Subscription creation</li><li>Resource provisioning</li></ul><h3>6. Social Media Has a Weird Problem: Millions of Followers</h3><p>Imagine a celebrity with <strong>50 million followers</strong>.</p><p>They post something.</p><p>You need to make that post appear in millions of feeds.</p><p>The naïve approach:</p><pre>Post<br> ↓<br>Find 50M followers<br> ↓<br>Update 50M feeds</pre><p>That is expensive.</p><p>But there’s another option.</p><h3>The hack: Fan-out on write</h3><p>When a user posts:</p><pre>Post<br> ↓<br>Fan-out Service<br> ↓<br>Follower Feed Storage<br> ↙   ↓   ↘<br>Feed A Feed B Feed C ...</pre><p>Then reading the feed becomes cheap.</p><p>The trade-off?</p><p><strong>Writes become expensive.</strong></p><h3>Alternative: Fan-out on read</h3><p>Instead:</p><pre>User opens feed<br>       ↓<br>Find followed users<br>       ↓<br>Fetch recent posts<br>       ↓<br>Merge<br>       ↓<br>Return feed</pre><p>Now reads become expensive.</p><p>So the real question isn’t:</p><blockquote><em>“Which pattern is correct?”</em></blockquote><p>It’s:</p><blockquote><strong><em>“Is my system read-heavy or write-heavy?”</em></strong></blockquote><p>Large systems often use a <strong>hybrid approach</strong>.</p><p>Normal users might use fan-out-on-write.</p><p>Extremely high-follower accounts might be handled differently.</p><p>That’s system design:</p><p><strong>not choosing a pattern, but choosing the right trade-off.</strong></p><h3>7. Uber-Style Location Queries: Don’t Search the Whole World</h3><p>Suppose a ride-sharing application needs to answer:</p><blockquote><strong><em>“Find drivers near me.”</em></strong></blockquote><p>You could theoretically search every driver.</p><p>Terrible idea.</p><p>If there are 5 million drivers, that’s a lot of unnecessary work.</p><h3>The hack: Geospatial indexing</h3><p>Divide the world into geographical cells.</p><p>For example:</p><pre>┌────┬────┬────┬────┐<br>│ A1 │ A2 │ A3 │ A4 │<br>├────┼────┼────┼────┤<br>│ B1 │ B2 │ B3 │ B4 │<br>├────┼────┼────┼────┤<br>│ C1 │ C2 │ C3 │ C4 │<br>└────┴────┴────┴────┘</pre><p>A driver belongs to a geographical region.</p><p>The user belongs to another.</p><p>Instead of: <strong>Search all drivers </strong>we do: <strong>Search nearby cells.</strong></p><p>Technologies and techniques can include:</p><ul><li>Geohash</li><li>H3</li><li>R-trees</li><li>Spatial databases</li><li>Redis geospatial indexes</li></ul><p>The architecture becomes:</p><pre>Driver Location<br>      ↓<br>Geo Index<br>      ↓<br>Nearby Cells<br>      ↓<br>Candidate Drivers<br>      ↓<br>Matching Algorithm</pre><p>The key lesson:</p><blockquote><strong><em>Index the dimension you search by.</em></strong></blockquote><p>If your primary query is geographical, your data structure should understand geography.</p><h3>8. Netflix-Style Streaming: Don’t Send the Same Video to Everyone</h3><p>Imagine a user watching a video on:</p><ul><li>A fast Wi-Fi connection</li><li>A 5G phone</li><li>A slow 4G network</li><li>An old television</li></ul><p>Sending the same 4K stream to everyone would be terrible.</p><h3>The hack: Adaptive bitrate streaming</h3><p>Create multiple versions:</p><pre>Video<br> ├── 4K<br> ├── 1080p<br> ├── 720p<br> ├── 480p<br> └── 360p</pre><p>The client continuously evaluates network conditions.</p><p>Then:</p><pre>Fast network → 1080p<br>Slow network → 480p<br>Very slow → 360p</pre><p>Protocols such as <strong>HLS</strong> and <strong>MPEG-DASH</strong> enable this type of adaptive streaming.</p><h3>Add a CDN</h3><p>Now combine adaptive streaming with edge caching:</p><pre>User<br> ↓<br>Nearest CDN<br> ↓<br>Cached Video Segment</pre><p>Instead of every user downloading content from the origin server.</p><p>This reduces:</p><ul><li>Latency</li><li>Origin bandwidth</li><li>Infrastructure load</li></ul><p>The larger lesson:</p><blockquote><strong><em>Move computation and data closer to the user whenever possible.</em></strong></blockquote><h3>9. When Your Database Becomes the Bottleneck, Don’t Immediately Buy a Bigger Database</h3><p>A common beginner reaction to database pressure is:</p><blockquote><em>“Let’s get a bigger machine.”</em></blockquote><p>Sometimes that works.</p><p>Eventually, it doesn’t.</p><p>Suppose your application has:</p><pre>10% Writes<br>90% Reads</pre><p>Using one database for both workloads can become inefficient.</p><h3>The hack: Read replicas</h3><p>Architecture:</p><pre>┌── Read Replica<br>              │<br>Application ──┼── Read Replica<br>              │<br>              └── Read Replica<br>                     ↑<br>                   Reads</pre><pre>Application<br>     ↓<br>Primary DB<br>     ↑<br>   Writes</pre><p>Writes go to the primary.</p><p>Reads are distributed across replicas.</p><p>Now your read capacity can scale horizontally.</p><h3>But there’s a catch</h3><p>Replication isn’t necessarily instantaneous.</p><p>Therefore:</p><p><strong>Read-after-write consistency</strong> can become important.</p><p>A user updates their profile.</p><p>Immediately afterward, they read it.</p><p>If the request goes to a lagging replica:</p><blockquote><em>“Where did my update go?”</em></blockquote><p>That’s the trade-off.</p><p>The lesson:</p><blockquote><strong><em>Scaling reads is easy. Scaling reads while preserving the right consistency guarantees is the real design problem.</em></strong></blockquote><h3>10. The Most Underrated Hack: Observe Everything</h3><p>This one isn’t a performance trick.</p><p>It’s a survival trick.</p><p>Your production system will fail.</p><p>Eventually.</p><p>The question is:</p><p><strong>Will you know why?</strong></p><p>Imagine this alert:</p><blockquote><strong><em>“Users are complaining that checkout is slow.”</em></strong></blockquote><p>Not useful.</p><p>Compare it with:</p><pre>Checkout p95 latency: 1.8s<br>Payment API p99: 4.2s<br>DB connection pool: 96%<br>Error rate: 3.7%<br>Kafka consumer lag: 280K<br>Region: ap-south-1</pre><p>Now you can investigate.</p><h3>The hack: The observability trio</h3><h4>1. Metrics</h4><p>“What is happening?”</p><p>Examples:</p><ul><li>CPU</li><li>Memory</li><li>Latency</li><li>Throughput</li><li>Error rate</li><li>Queue depth</li></ul><h4>2. Logs</h4><p>“What happened?”</p><p>Useful for detailed events and debugging.</p><h4>3. Traces</h4><p>“Where did the request spend its time?”</p><p>For a distributed request:</p><pre>Client<br> ↓<br>API Gateway<br> ↓<br>Order Service<br> ↓<br>Payment Service<br> ↓<br>Database</pre><p>Distributed tracing can show exactly where latency entered the system.</p><h3>Then add SLOs</h3><p>Instead of alerting on everything, define meaningful objectives.</p><p>For example:</p><blockquote><strong><em>99.9% of checkout requests should complete successfully within X milliseconds.</em></strong></blockquote><p>Now alerts become connected to user experience.</p><p>The fundamental lesson:</p><blockquote><strong><em>You can’t reliably operate a system you can’t observe.</em></strong></blockquote><h3>Putting the 10 Hacks Together</h3><p>The interesting thing is that these techniques aren’t isolated.</p><p>A large production system might combine many of them.</p><p>For example, imagine an e-commerce flash sale:</p><pre>Users<br>                     ↓<br>                  CDN<br>                     ↓<br>               API Gateway<br>                     ↓<br>               Rate Limiter<br>                     ↓<br>                Load Balancer<br>                     ↓<br>              Application Tier<br>                 ↙       ↘<br>              Cache      Queue<br>                ↓          ↓<br>                DB       Workers<br>                 ↓          ↓<br>              Replicas   Events<br>                           ↓<br>                    Notifications</pre><p>And surrounding everything:</p><pre>Metrics + Logs + Traces + Alerts</pre><p>Suddenly, system design stops looking like:</p><p><strong>“Which database should I use?”</strong></p><p>And starts looking like:</p><blockquote><strong><em>“How do I control traffic, isolate failures, minimize latency, scale independently, and preserve correctness?”</em></strong></blockquote><p>That’s the real game.</p><h3>The 10 Hacks, in One Table</h3><p>ProblemHackMain BenefitMassive synchronized traffic<strong>Queue + Rate Limiting</strong>Absorb spikesPredictable traffic surges<strong>Predictive Autoscaling</strong>Scale before demandRepeated reads<strong>Caching</strong>Reduce DB loadNon-critical work<strong>Async Processing</strong>Faster responsesNetwork retries<strong>Idempotency Keys</strong>Prevent duplicatesHuge social graphs<strong>Fan-out Strategies</strong>Scale feedsNearby-location queries<strong>Geospatial Indexing</strong>Fast location searchVideo delivery<strong>Adaptive Bitrate + CDN</strong>Lower latencyRead-heavy workloads<strong>Read Replicas</strong>Horizontal read scalingProduction complexity<strong>Observability</strong>Faster diagnosis</p><h3>Final Thought: System Design Is Mostly About Trade-offs</h3><p>There is no magical architecture.</p><p>Queues introduce latency.</p><p>Caching introduces staleness.</p><p>Replication introduces consistency challenges.</p><p>Sharding introduces operational complexity.</p><p>Asynchronous processing introduces eventual consistency.</p><p>Microservices introduce distributed-system problems.</p><p>Every optimization has a price.</p><p>And that’s precisely what makes system design interesting.</p><p>The best system designer isn’t the person who knows the most technologies.</p><p>It’s the person who can look at a problem and say:</p><blockquote><strong><em>“This is the bottleneck. This is the failure mode. This is the trade-off. And this is the simplest architecture that solves it.”</em></strong></blockquote><p>So the next time you see:</p><p><strong>10:00 AM Tatkal.</strong></p><p><strong>A flash sale.</strong></p><p><strong>A UPI payment.</strong></p><p><strong>A food-delivery driver moving across your screen.</strong></p><p><strong>A Netflix video switching quality automatically.</strong></p><p>Don’t just see an app.</p><p>Look underneath.</p><p>There’s probably a <strong>queue, cache, replica, index, stream, retry mechanism, or carefully designed trade-off</strong> quietly keeping the whole thing alive.</p><p>And that’s where system design gets really fun. 🚀</p><p><em>Note: The company examples in this article are used as real-world scenarios to illustrate system-design principles. The architectural diagrams are conceptual and should not be interpreted as descriptions of proprietary internal implementations.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=008c0338b169" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Being Busy Doesn’t Mean You’re Productive]]></title>
            <link>https://adosi12.medium.com/why-being-busy-doesnt-mean-you-re-productive-88608a8a5e41?source=rss-7b54740e75bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/88608a8a5e41</guid>
            <category><![CDATA[work-life-balance]]></category>
            <category><![CDATA[self-improvement]]></category>
            <category><![CDATA[personal-development]]></category>
            <category><![CDATA[time-management]]></category>
            <category><![CDATA[productivity]]></category>
            <dc:creator><![CDATA[Aditi Dosi]]></dc:creator>
            <pubDate>Sun, 23 Aug 2026 15:57:03 GMT</pubDate>
            <atom:updated>2026-08-23T15:57:03.615Z</atom:updated>
            <content:encoded><![CDATA[<p><em>You can spend 10 hours working and still not make meaningful progress.</em></p><p>We all know that feeling. You wake up with a long to-do list, check your emails, reply to a few messages, attend meetings, jump between tasks, and somehow spend the entire day “working.” By evening, you’re exhausted. You close your laptop and think, <em>“I had such a busy day.”</em></p><p>But then comes the uncomfortable question: <strong>What did I actually accomplish?</strong></p><p>Sometimes, the answer isn’t very satisfying.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/856/0*TFIwv7XyYW8wl1RV" /></figure><p>We have quietly started confusing <strong>being busy with being productive</strong>. If our calendar is full and our notifications keep popping up, we feel like we’re making progress. But activity and progress aren’t the same thing.</p><p>Think about a hamster running on a wheel. It’s moving constantly, putting in effort, probably working harder than most of us on a Monday morning. But it’s not going anywhere.</p><p>Our workdays can look surprisingly similar.</p><p>Emails, meetings, WhatsApp messages, Slack notifications, calls, spreadsheets, quick tasks and “just five minutes” on social media. The wheel keeps spinning. And because we’re constantly doing something, we convince ourselves that we’re being productive.</p><p>I noticed this in my own work too.</p><p>I would start the morning with a clear plan: <em>“I’m going to spend the first two hours working on the important project.”</em> Then I’d think, <em>“Let me just quickly check my email first.”</em></p><p>That five-minute email check would become fifteen minutes. One email needed a response. That response created another question. Then I’d check a message. Then another notification would appear. Before I knew it, an hour had disappeared.</p><p>And the funny thing was, I had actually done a lot.</p><p>I had answered emails. I had replied to messages. I had cleared notifications.</p><p>I had been <strong>busy</strong>.</p><p>But the important project hadn’t moved an inch.</p><p>That’s when I started noticing something uncomfortable: <strong>the easiest tasks often make us feel the most productive.</strong></p><p>There is an instant reward when you answer an email. Done. Another reward when you clear a notification. Done. Another when you finish a small task. Done.</p><p>But meaningful work rarely works like that.</p><p>Writing an article can take hours. Learning a new skill can take weeks. Building a business can take months. Improving your health requires consistency. Creating something valuable often means sitting with uncertainty and not knowing whether your effort will pay off.</p><p>There isn’t always a little green checkmark telling you that you’re making progress.</p><p>So we escape toward easier things.</p><p>We organize our files. We clean our inbox. We make another to-do list. We watch another productivity video. We tell ourselves we’re preparing to work.</p><p>Sometimes, <strong>procrastination doesn’t look like laziness. It looks incredibly productive.</strong></p><p>That’s the trap.</p><p>We tend to prioritize whatever is urgent, visible or noisy. A notification demands attention immediately. An email sits in the inbox looking important. A meeting appears on the calendar.</p><p>Meanwhile, the thing that could actually change our career, business or life quietly waits in the background.</p><p>This is where I think productivity needs a different definition.</p><p>Productivity isn’t about asking, <strong>“How much did I do today?”</strong></p><p>It’s about asking, <strong>“Did the things I did today actually matter?”</strong></p><p>Those are two very different questions.</p><p>You can work for ten hours and make very little progress. You can also spend three focused hours on the right problem and move something important forward.</p><p>The goal isn’t to fill every empty space in your day.</p><p>The goal is to protect your attention for the things that deserve it.</p><p>One question has helped me with this: <strong>“If I could accomplish only one meaningful thing today, what would it be?”</strong></p><p>Not five things. Not ten.</p><p>One.</p><p>Because once you know what matters most, it’s much easier to recognize what doesn’t.</p><p>That email can probably wait thirty minutes. That notification can wait. That meeting might not need you. That perfectly formatted presentation might not need another hour of tweaking.</p><p>We often forget that <strong>urgency isn’t the same as importance.</strong></p><p>And there’s another question worth asking: <strong>“What am I avoiding?”</strong></p><p>This one is harder.</p><p>Sometimes we’re busy because we’re avoiding the difficult task.</p><p>We answer emails because writing the report feels overwhelming. We organize our desk because starting the project feels uncomfortable. We scroll through our phones because taking the first step toward something important feels risky.</p><p>Small tasks feel safe.</p><p>Important tasks often don’t.</p><p>What if the idea fails? What if people don’t like what I create? What if I’m not good enough?</p><p>It’s much easier to check your inbox than to confront those questions.</p><p>Maybe that’s why being busy can sometimes feel strangely comfortable. <strong>Busy gives us the feeling of progress without forcing us to face uncertainty.</strong></p><p>The irony is that the work that matters most is often the work we find easiest to postpone.</p><p>So tomorrow, try something simple.</p><p>Before opening your inbox or checking your notifications, take five minutes and write down the one thing that would make your day feel worthwhile.</p><p>Then give that task your best hour.</p><p>No multitasking. No “quick” email check. No switching between tabs every five minutes.</p><p>Just one thing.</p><p>You may discover that an hour of focused work can be more valuable than an entire day of scattered attention.</p><p>We’ve built a culture where being busy is almost something to be proud of. We say things like, <em>“I’m slammed,” “I’ve been crazy busy,”</em> or <em>“I’ve had back-to-back meetings,”</em> as if exhaustion itself is evidence of achievement.</p><p>But a full calendar isn’t necessarily a successful calendar.</p><p>A long to-do list isn’t necessarily an impressive one.</p><p>And working late doesn’t automatically mean you’re productive.</p><p>At the end of the year, you probably won’t remember most of the emails you answered or the meetings you attended.</p><p>You’ll remember the project you finally started. The skill you learned. The book you wrote. The business you built. The relationship you invested in. The person you became.</p><p>That’s the difference between <strong>spending your time and investing your time.</strong></p><p>So the next time you catch yourself saying, <em>“I’ve been so busy,”</em> pause for a second and ask:</p><p><strong>“Busy doing what?”</strong></p><p>Because sometimes the most productive thing you can do isn’t adding another task to your list.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/906/0*822Y4iO7JdGhNPmU" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/611/1*JqnofQcnNKcQqA8cbzZyFg.png" /></figure><blockquote><strong>It’s removing the noise and focusing on what actually matters.</strong></blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=88608a8a5e41" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Failure Isn’t the End: What If It’s Just Version 1.0?]]></title>
            <link>https://adosi12.medium.com/failure-isnt-the-end-what-if-it-s-just-version-1-0-505f1fa03641?source=rss-7b54740e75bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/505f1fa03641</guid>
            <category><![CDATA[life]]></category>
            <category><![CDATA[consistency]]></category>
            <category><![CDATA[resilience]]></category>
            <category><![CDATA[life-lessons]]></category>
            <category><![CDATA[mindset]]></category>
            <dc:creator><![CDATA[Aditi Dosi]]></dc:creator>
            <pubDate>Sat, 22 Aug 2026 07:34:50 GMT</pubDate>
            <atom:updated>2026-08-22T07:34:50.840Z</atom:updated>
            <content:encoded><![CDATA[<p>There is a strange little voice that visits us whenever something goes wrong.</p><p><em>“I failed.”</em></p><p>We hear it after a rejected job application, a business idea that goes nowhere, an exam we didn’t pass, a relationship that falls apart, or even a simple presentation that goes horribly off the rails.</p><p>And once that voice starts talking, it can turn a small mistake into a full-blown identity crisis.</p><p>But what if we changed the sentence?</p><p>Instead of saying, <strong>“I failed,”</strong> what if we said, <strong>“That attempt didn’t work”?</strong></p><p>It sounds like a small difference, but it can change everything.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*0mBvfTnQo1vvcvNL" /></figure><h3>Failure Has a Bad Reputation</h3><p>We are taught to think of success as a straight road: work hard, make good decisions, achieve the goal.</p><p>Real life, of course, rarely gets the memo.</p><p>The road is full of potholes, wrong turns, detours, and the occasional dead end.</p><p>Thomas Edison famously experimented thousands of times before developing a practical electric light bulb. Whether we look at his story, a scientist testing a hypothesis, or a child learning to ride a bicycle, the pattern is the same:</p><p><strong>Try. Observe. Adjust. Try again.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*5q1X_E6cE5sQdWCc" /></figure><p>That is not failure.</p><p>That is iteration.</p><p>The word <em>iteration</em> sounds rather technical, but the idea is beautifully simple. You do something, learn from what happened, make a change, and give it another shot.</p><p>In other words, you don’t throw away the whole recipe because the first pancake came out looking like a map of Australia.</p><p>You adjust the batter.</p><h3>The Experimenter’s Mindset</h3><p>Imagine two people applying for their dream job.</p><p>The first person doesn’t get selected.</p><p>They think:</p><blockquote><em>“I’m obviously not good enough.”</em></blockquote><p>The second person thinks:</p><blockquote><em>“Interesting. My approach didn’t produce the result I wanted. What can I learn?”</em></blockquote><p>The first person closes the door.</p><p>The second person looks for the next door.</p><p>This is the power of thinking like an experimenter.</p><p>Scientists don’t usually take a failed experiment personally. If a hypothesis doesn’t work, they don’t stare dramatically into the rain and whisper, <em>“Perhaps I am a terrible scientist.”</em></p><p>They ask better questions.</p><blockquote><strong>What happened?</strong></blockquote><blockquote><strong>Why might it have happened?</strong></blockquote><blockquote><strong>What should I change next time?</strong></blockquote><p>We can borrow that attitude in everyday life.</p><p>A failed interview can teach you which questions you weren’t prepared for.</p><p>A disappointing project can reveal where your planning went wrong.</p><p>A business idea that didn’t take off can show you what customers actually want.</p><p>A difficult conversation can teach you how you communicate under pressure.</p><p>The mistake becomes data.</p><p>And data is useful.</p><h3>Don’t Turn One Bad Chapter Into the Whole Book</h3><p>One of the biggest traps of failure is overgeneralization.</p><blockquote>You make one mistake and suddenly your brain produces a greatest-hits album of everything you’ve ever done wrong.</blockquote><blockquote><em>I messed up this presentation.</em></blockquote><blockquote>becomes:</blockquote><blockquote><em>I’m terrible at public speaking.</em></blockquote><blockquote>Then:</blockquote><blockquote><em>I’m not good at my job.</em></blockquote><blockquote>Then:</blockquote><blockquote><em>I’ll never succeed.</em></blockquote><p>Before you know it, you’ve gone from one bad Tuesday to predicting the next twenty years of your life.</p><p>That’s quite a leap.</p><p>A healthier approach is to separate <strong>what happened</strong> from <strong>who you are</strong>.</p><p>You didn’t <em>become</em> a failure.</p><p>You experienced an unsuccessful outcome.</p><p>There is a world of difference between those two statements.</p><p>Your mistake is an event, not an identity.</p><h3>The Anxiety Shrinks When the Lesson Grows</h3><p>Failure often creates anxiety because we imagine it is permanent.</p><p>But when we treat mistakes as experiments, the question changes.</p><blockquote>Instead of:</blockquote><blockquote><strong>“What if I fail?”</strong></blockquote><blockquote>we begin asking:</blockquote><blockquote><strong>“What will I learn if this doesn’t work?”</strong></blockquote><p>That question gives us room to breathe.</p><p>It doesn’t make disappointment disappear. Nobody enjoys rejection or watching months of effort fall flat.</p><p>But it gives disappointment a purpose.</p><p>And purpose makes difficult experiences easier to carry.</p><p>Think of a GPS.</p><p>When you take a wrong turn, it doesn’t scream, <em>“You are a terrible driver!”</em></p><p>It simply recalculates.</p><p><strong>Recalculating.</strong></p><p>There is something deeply comforting about that word.</p><p>Maybe we should do the same.</p><h3>Your Life Doesn’t Need to Be Perfect on the First Draft</h3><p>We sometimes expect ourselves to have everything figured out immediately.</p><p>The perfect career.</p><p>The perfect relationship.</p><p>The perfect routine.</p><p>The perfect financial plan.</p><p>The perfect version of ourselves.</p><p>But nobody writes a brilliant first draft.</p><p>Writers edit.</p><p>Designers revise.</p><p>Entrepreneurs pivot.</p><p>Athletes practice.</p><p>Musicians repeat the same difficult passage until their fingers finally understand what their brain has been asking them to do.</p><p>So why should our lives be any different?</p><p>You are allowed to be a work in progress.</p><p>You are allowed to change your mind.</p><p>You are allowed to discover that something you wanted five years ago isn’t what you want today.</p><p>You are allowed to say, <em>“I tried that. It didn’t work for me. Let’s try something else.”</em></p><p>That’s not giving up.</p><p>Sometimes, it’s growing up.</p><h3>Three Questions to Ask After Something Goes Wrong</h3><p>The next time life throws you a curveball, resist the urge to immediately judge yourself.</p><p>Take a breath and ask:</p><p><strong>1. What actually happened?</strong></p><p>Stick to the facts instead of the dramatic story your mind may be creating.</p><p><strong>2. What did this teach me?</strong></p><p>Look for information rather than blame.</p><p><strong>3. What will I do differently next time?</strong></p><p>Turn the lesson into an action.</p><p>These three questions can turn a painful experience into a stepping stone.</p><p>And yes, sometimes the stepping stone is slippery. That’s life. But at least you’re moving forward.</p><h3>The Lesson</h3><p>Perhaps the goal isn’t to build a life where nothing goes wrong.</p><p>That would be impossible.</p><p>Perhaps the goal is to become someone who knows what to do <strong>when</strong> things go wrong.</p><p>Failure can close a door, but it can also point toward a better one.</p><p>A rejected application might push you toward a role you never considered.</p><p>A failed project might reveal a skill you need to develop.</p><p>A wrong decision might teach you exactly what matters to you.</p><blockquote>The trick is to stop asking, <strong>“Why am I such a failure?”</strong></blockquote><blockquote>Ask instead:</blockquote><blockquote><strong>“What is this experience trying to teach me?”</strong></blockquote><p>Because sometimes life doesn’t hand us a finished masterpiece.</p><p>Sometimes it hands us a rough draft, a few crossed-out lines, and a blank page.</p><p>And perhaps that’s not a punishment.</p><p>Perhaps it’s an invitation.</p><h3><strong>Edit. Learn. Try again.</strong></h3><p>Your last attempt doesn’t have to be your final version.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=505f1fa03641" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Three Years In: Lessons Beyond the Job Description]]></title>
            <link>https://adosi12.medium.com/three-years-in-lessons-beyond-the-job-description-1fa4bd7359b5?source=rss-7b54740e75bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/1fa4bd7359b5</guid>
            <category><![CDATA[life-lessons]]></category>
            <category><![CDATA[work-life-balance]]></category>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[life]]></category>
            <category><![CDATA[corporate-culture]]></category>
            <dc:creator><![CDATA[Aditi Dosi]]></dc:creator>
            <pubDate>Sat, 01 Aug 2026 15:13:18 GMT</pubDate>
            <atom:updated>2026-08-01T15:13:18.624Z</atom:updated>
            <content:encoded><![CDATA[<p>Completing three years at my first workplace this July has been a milestone I&#39;ll always be grateful for. It gave me the opportunity to build my career from scratch, work alongside incredibly talented people, and solve problems that shaped me into a better engineer.</p><p>But looking back, the biggest lessons weren&#39;t about writing better code or designing better systems. They came from <strong>quietly observing people.</strong></p><p>When I joined, I imagined the workplace as a place where everyone worked toward the same goal. Collaboration was what I expected to see every day.</p><p>Over time, I realized that <strong>another dynamic quietly exists alongside collaboration.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TPwecGfT0VNzL1CCDgRy_w.png" /><figcaption>Lessons that could never have been written in a job description</figcaption></figure><p>What surprised me wasn&#39;t that people compete. Competition is natural when everyone is trying to grow. What surprised me was <strong>how silent that competition can be.</strong></p><p>I&#39;ve seen senior colleagues who looked like the best of friends. They would discuss work, personal life, laugh together over coffee, help each other during the day, and genuinely seem inseparable.</p><p>But the moment <strong>recognition, visibility, or appreciation</strong> entered the picture, the atmosphere sometimes changed in subtle ways.</p><p>I noticed people becoming uncomfortable when someone else&#39;s work received praise. I saw work that juniors had spent weeks building being presented by someone else because they happened to be leading the project. I watched people become exceptionally good at making sure their work reached the right audience. I also noticed others quietly building stronger relationships with managers—not because it was wrong, but because they understood that <strong>visibility often matters just as much as performance.</strong></p><p>None of this made me think people were bad.</p><p>If anything, it helped me understand that everyone is carrying their own ambitions, insecurities, and career goals. People simply learn different ways to move forward.</p><p>That realization changed the way I looked at the workplace. <strong>Friendly conversations don&#39;t always mean aligned interests</strong>, and that&#39;s completely normal in an environment where everyone is trying to grow.</p><p>Thankfully, that wasn&#39;t the only lesson.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0jSs64LJzz4HowiLoPWqTw.png" /></figure><p>Among all the people I&#39;ve met over the last three years, <strong>a handful of people became genuine mentors.</strong> They&#39;re the ones who invested time in teaching, answered questions patiently, celebrated my progress, and wanted nothing in return except to see me grow.</p><p>Those are the people who leave a lasting impact on your career. Long after projects are forgotten, <strong>you remember the people who believed in you</strong> when you were still finding your feet.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4Ya4MxCJbr0Rr3f2nmughw.png" /></figure><p>The biggest lesson, however, came from looking inward rather than outward.</p><p>I realized I couldn&#39;t control how visible my work would be, how others perceived it, or how different people approached their careers. <strong>But I could control how much I learned.</strong></p><p>That&#39;s why every year I&#39;ve tried to push myself outside my regular work by participating in hackathons. This year became especially memorable when <strong>our team finished in the Top 3 among 180 teams.</strong></p><p>That achievement reminded me that the most valuable investment is still your own growth. Skills stay with you. Curiosity stays with you. The ability to solve difficult problems stays with you. <strong>Those things don&#39;t disappear when projects end or organizations change.</strong></p><p>Looking back, these three years haven&#39;t made me cynical.</p><p><strong>They&#39;ve made me more aware.</strong></p><p>I&#39;ve learned that ambition and kindness can exist in the same person. That workplaces are more nuanced than they first appear. That recognition doesn&#39;t always tell the complete story. And that genuine mentors are far rarer—and far more valuable—than I ever imagined.</p><p>Most importantly, I&#39;ve learned that <strong>the best response isn&#39;t to get caught up in every workplace dynamic.</strong> It&#39;s to keep learning, keep building, and let your work quietly speak for itself.</p><blockquote>I&#39;m grateful for where my journey began, for every person who helped shape it, and for the lessons that could never have been written in a job description.</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1fa4bd7359b5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Fandom Trap: How Watching Them Kills Your Motivation]]></title>
            <link>https://adosi12.medium.com/the-fandom-trap-how-watching-them-kills-your-motivation-e346f5bc3046?source=rss-7b54740e75bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/e346f5bc3046</guid>
            <category><![CDATA[developer]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[creativity]]></category>
            <category><![CDATA[motivation]]></category>
            <category><![CDATA[dreams]]></category>
            <dc:creator><![CDATA[Aditi Dosi]]></dc:creator>
            <pubDate>Sun, 12 Jul 2026 13:12:13 GMT</pubDate>
            <atom:updated>2026-07-12T13:12:13.175Z</atom:updated>
            <content:encoded><![CDATA[<p>The Fandom Trap: How Watching Them Kills Your Motivation</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q5ME9GY2WnverEEXl9nh6Q.jpeg" /><figcaption>Stop Dreaming, Start Doing (Seriously, Put the Phone Down)</figcaption></figure><h3>The Ultimate Fan, The Ultimate Benchwarmer</h3><p>Think about a typical evening. You finally close your IDE, log off from your 9-to-5, and promise yourself you’ll spend an hour mastering Data Structures &amp; Algorithms or building that side project API. You’ve got the architecture planned out in your head. It’s beautiful. It’s scalable.</p><p>But then, the doom-scrolling loop takes over, and your personal roadmap suffers a major merge conflict.</p><p><em>Suddenly, you’ve spent two hours:</em></p><ul><li><strong><em>Deep-diving into Instagram Reels or YouTube Shorts</em></strong><em> instead of diving into your codebase.</em></li><li><strong><em>Binge-watching a full season of a new show</em></strong><em> because “just one more episode” has higher priority than your actual life goals.</em></li><li><strong><em>Fiercely defending your favorite online creator, actor, or reality TV star</em></strong><em> in the comment section like you’re debugging production code on a Friday night.</em></li></ul><p>There is absolutely nothing wrong with entertainment. We all need to unwind and flush our mental cache! But here is the plot twist we usually ignore: while we are spending our precious RAM cheering on our favorite personalities and celebrating their massive success, they are fulfilling <em>their</em> dreams. They are successfully deploying their life goals to production.</p><p><strong>Meanwhile, our own dreams are stuck in the backlog, waiting for a sprint that never actually gets planned.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0N_PrTgO3PbTbDfWKcidPg.jpeg" /><figcaption>Visualize the Progress</figcaption></figure><p>This whimsical illustration captures the moment we abandon the couch (and the seductive pull of the scrolling phone) and actually <em>take our dreams for a walk</em>. Notice how the abstract, swirling ‘dream cloud’ (now featuring the successful writer, coder, and presenter scenes) is actively walking <em>with</em> the determined figure on the path. We’ve even added a handy, humorous progress bar because, let’s be real, even 1% committed is a lot better than 0%! It’s all about making that tiny, daily effort that keeps the momentum going.</p><h3>Moving the Needle, One Inch (or 20 Minutes) at a Time</h3><p>Living a dream doesn’t mean you have to quit your day job tomorrow, burn your bridges, and work 18 hours straight. It just means checking in with yourself every day and asking a simple, profound question: <strong>“Did I give my dream at least 20 minutes today, or did I give it all to a relentless algorithm that knows I have a weakness for cat videos?”</strong></p><p>Instead of just <em>thinking</em> about the glorious, cinematic end goal, we have to start doing the unglamorous, often tedious, small steps that actually build it. It’s about less passive consuming and a lot more active creating.</p><p>So, let’s make a deal: let’s close the 50 extra tabs we have open “for research,” put the phone on silent (because the algorithm <em>will</em> tempt us with another perfect recommendation), and actually take our dreams out for a walk today. Because no one else is going to build them for us. Not even AI.</p><blockquote>“Your favorite creators are busy deploying their dreams to production. It’s time to stop debugging their lives and start writing the code for your own.”</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e346f5bc3046" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Welcome to the world of TypeScript]]></title>
            <link>https://adosi12.medium.com/welcome-to-the-world-of-typescript-08b83f2d2a9f?source=rss-7b54740e75bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/08b83f2d2a9f</guid>
            <category><![CDATA[ui]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[typescript]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[javascript]]></category>
            <dc:creator><![CDATA[Aditi Dosi]]></dc:creator>
            <pubDate>Sat, 16 May 2026 12:19:00 GMT</pubDate>
            <atom:updated>2026-05-16T12:19:00.976Z</atom:updated>
            <content:encoded><![CDATA[<p>Think of it as JavaScript with superpowers — it catches your mistakes before your code even runs, saving you from countless debugging headaches.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aWOga6bVm_CNOL1VEVMkzA.png" /></figure><h4>1. Let TypeScript Do the Heavy Lifting (Inference)</h4><pre>// ❌ Overkill: You don&#39;t need to explicitly type this<br>const username: string = &quot;Alice&quot;;<br>//  Excellent: TypeScript already knows this is a string<br>const username = &quot;Alice&quot;;</pre><h4>2. Learn the Difference: interface vs. type</h4><pre>// Best for object structures<br>interface User {<br> id: number;<br> name: string;<br>}<br>// Best for combining or restricting values (Union types)<br>type Status = &quot;pending&quot; | &quot;approved&quot; | &quot;rejected&quot;;</pre><h4>3. Avoid `any` Like the Plague</h4><pre>// ❌ Avoid this<br>let data: any = &quot;Hello&quot;;<br>data.helloWorld(); // Crash at runtime! TypeScript won&#39;t warn you.<br>// Do this instead<br>let data: unknown = &quot;Hello&quot;;<br>if (typeof data === &quot;string&quot;) {<br> console.log(data.toUpperCase()); // Safe!<br>}</pre><h4>4. Embrace Optional Properties</h4><pre>interface Product {<br> name: string;<br> price: number;<br> description?: string; // This property is optional<br>}<br>const shirt: Product = { name: &quot;T-Shirt&quot;, price: 20 }; // Perfectly valid!</pre><h4>5. Master the Shortcut: tsconfig.json</h4><p>Your tsconfig.json file is the brain of your TypeScript project. To get the most out of TypeScript&#39;s protective shield, always ensure <strong>strict mode</strong> is turned on.</p><pre>{<br> &quot;compilerOptions&quot;: {<br> &quot;strict&quot;: true<br> }<br>}</pre><p><strong>Why?</strong> It enables a suite of checks (like strictNullChecks) that prevent the dreaded <em>&quot;Cannot read property of undefined&quot;</em> errors. It might feel strict at first, but it writes much safer code.</p><h4>6. Use the Record Utility for Dynamic Objects</h4><pre>// A dictionary where the keys are strings (usernames) and the values are numbers (scores)<br>const userScores: Record&lt;string, number&gt; = {<br> alice: 95,<br> bob: 88,<br>};</pre><p>If you are looking at a <strong>hybrid codebase</strong> or migrating an existing JavaScript project over to TypeScript. This is an incredibly common scenario!</p><p>Because you are using both .js and .ts files, you need to configure your environment so that JavaScript and TypeScript can coexist peacefully while relying on your package.json to manage dependencies and scripts.</p><blockquote>Here is how to set up and manage a mixed project effectively.</blockquote><h4>1. The Key Configuration: tsconfig.json</h4><p>Even though your project uses package.json for managing packages, TypeScript absolutely requires a tsconfig.json file in the root directory to understand how to handle both file types.</p><p>To allow TypeScript to read and compile your JavaScript files alongside your TypeScript files, you must enable two specific flags: <strong>allowJs</strong> and <strong>checkJs</strong>.</p><pre>{<br>  &quot;compilerOptions&quot;: {<br>    &quot;target&quot;: &quot;ES2022&quot;,<br>    &quot;module&quot;: &quot;CommonJS&quot;,<br>    <br>    // 💡 CRITICAL FOR MIXED PROJECTS:<br>    &quot;allowJs&quot;: true,      // Allows JavaScript files to be compiled<br>    &quot;checkJs&quot;: false,     // Set to true if you want TS to report errors in .js files<br>    <br>    &quot;outDir&quot;: &quot;./dist&quot;,   // Where the compiled JS files will go<br>    &quot;strict&quot;: true<br>  },<br>  &quot;include&quot;: [&quot;src/**/*&quot;] // Tells TS to look at all files inside the src folder<br>}</pre><h3>2. Setting Up your package.json</h3><p>Your package.json file needs the correct dependencies and scripts to build your mixed project.</p><h3>Step A: Install the Required Dependencies</h3><p>You need typescript installed as a development dependency. Run this in your terminal:</p><p>Bash</p><pre>npm install --save-dev typescript @types/node</pre><h3>Step B: Add a Build Script</h3><p>In your package.json, add a script to trigger the TypeScript compiler. When you run this script, TypeScript will take <strong>both</strong> your .ts and .js files and output clean, standard JavaScript into your destination folder (like ./dist).</p><p>JSON</p><pre>{<br>  &quot;name&quot;: &quot;my-mixed-project&quot;,<br>  &quot;version&quot;: &quot;1.0.0&quot;,<br>  &quot;main&quot;: &quot;dist/index.js&quot;, // Point this to your compiled output main file<br>  &quot;scripts&quot;: {<br>    &quot;build&quot;: &quot;tsc&quot;,<br>    &quot;start&quot;: &quot;node dist/index.js&quot;<br>  },<br>  &quot;devDependencies&quot;: {<br>    &quot;@types/node&quot;: &quot;^20.0.0&quot;,<br>    &quot;typescript&quot;: &quot;^5.0.0&quot;<br>  }<br>}</pre><h3>3. How to Import Files Between .js and .ts</h3><p>When you have a mixed project, you can easily share code between files, but you have to pay attention to how you import them.</p><h3>Importing JS into TS</h3><p>If you have a file named utils.js and you want to use it inside app.ts:</p><p>JavaScript</p><pre>// utils.js<br>function greet(name) {<br>  return `Hello, ${name}`;<br>}<br>module.exports = { greet };</pre><p>TypeScript</p><pre>// app.ts<br>// TypeScript will automatically read utils.js because &quot;allowJs&quot; is true<br>import { greet } from &#39;./utils&#39;; </pre><pre>console.log(greet(&quot;TypeScript&quot;));</pre><blockquote><em>⚠️ </em><strong><em>Note on Types:</em></strong><em> Because </em><em>utils.js doesn&#39;t have explicit types, TypeScript will treat </em><em>greet as having an </em><em>any type unless you provide a definition for it, or use JSDoc comments inside the JS file.</em></blockquote><p>Share with me if you are building any project using typescript :) Will be happy to hear.</p><p><em>Aditi…</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=08b83f2d2a9f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why a person should shift on blockchain transaction mode ?]]></title>
            <link>https://adosi12.medium.com/why-a-person-should-shift-on-blockchain-transaction-mode-f338d224530f?source=rss-7b54740e75bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/f338d224530f</guid>
            <category><![CDATA[smart-contracts]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[escrow]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[blockchain-technology]]></category>
            <dc:creator><![CDATA[Aditi Dosi]]></dc:creator>
            <pubDate>Sun, 26 Apr 2026 07:43:50 GMT</pubDate>
            <atom:updated>2026-04-26T07:43:50.305Z</atom:updated>
            <content:encoded><![CDATA[<p>For a regular person, the “why” usually boils down to three technical advantages that traditional banks (centralized systems) struggle to match:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*eCCFHaYEPjgC-Hsh" /></figure><ul><li><strong>Settlement Speed (T+0):</strong> In traditional banking, a cross-border transfer can take 3–5 days to “settle” because multiple banks have to talk to each other. On a blockchain, the transaction and the settlement happen <strong>simultaneously</strong>. Once the block is mined, the money is there.</li><li><strong>Programmability (Smart Contracts):</strong> You can’t tell a 10-dollar bill to “only spend yourself if the weather is 30°C.” You <em>can</em> do that with blockchain. This allows for automated escrow, streaming salaries (getting paid by the second), and automated refunds.</li><li><strong>24/7/365 Availability:</strong> Blockchains don’t have “bank holidays” or “working hours.” The network is global and always on.</li><li><strong>Self-Sovereignty:</strong> You are your own bank. No central authority can “freeze” your account because of a technical glitch or a policy change, provided you manage your private keys securely.</li></ul><h3>Live Projects Used by the Public (2026)</h3><p>Blockchain has moved past the “experimental” phase. Here are verified applications being used by the public and major institutions today.</p><p>Based on the table provided, here is a breakdown of those live projects categorized for your article:</p><h3>Institutional Finance &amp; Investment</h3><ul><li><strong>JPM Coin (JPMorgan):</strong> Enables large corporations to move billions of dollars across international accounts instantly, 24/7, bypassing traditional banking delays.</li><li><strong>BlackRock BUIDL:</strong> A tokenized fund on the Ethereum blockchain that provides institutional investors with instant liquidity and yield on US Treasuries.</li></ul><h3>Supply Chain &amp; Global Trade</h3><ul><li><strong>Tracr (De Beers):</strong> Secures the diamond industry by tracking over 25% of global production by value, ensuring stones are ethically sourced and “conflict-free.”</li><li><strong>IBM Food Trust:</strong> Utilized by retail giants like Walmart and Carrefour to trace the origin of produce and meat, reducing the time to identify foodborne illness sources from days to minutes.</li></ul><h3>Digital Economy &amp; Government</h3><ul><li><strong>Magic Eden:</strong> A high-volume NFT marketplace where millions of users buy, sell, and trade digital collectibles and in-game assets.</li><li><strong>Estonia’s X-Road:</strong> A national-scale infrastructure using blockchain-like technology to secure the identity and health records of 99% of Estonia’s citizens.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f338d224530f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why shift to Blockchain ?]]></title>
            <link>https://adosi12.medium.com/why-shift-to-blockchain-9926b0b4e894?source=rss-7b54740e75bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/9926b0b4e894</guid>
            <category><![CDATA[solidity]]></category>
            <category><![CDATA[dapps]]></category>
            <category><![CDATA[blockchain]]></category>
            <category><![CDATA[smart-contracts]]></category>
            <category><![CDATA[celestia]]></category>
            <dc:creator><![CDATA[Aditi Dosi]]></dc:creator>
            <pubDate>Sun, 26 Apr 2026 07:33:07 GMT</pubDate>
            <atom:updated>2026-04-26T07:33:07.837Z</atom:updated>
            <content:encoded><![CDATA[<p>Building an end-to-end blockchain application (dApp) is a fantastic way to master the integration of smart contracts with modern web frameworks.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/1*GugYmacC0Pr8I8a3fKBefA.png" /></figure><p>Here are four <strong>distinct directions</strong> ranging from supply chain to finance, each designed to showcase a full-stack technical grasp.</p><h3>1. Decentralized Milestone-Based Freelance Marketplace</h3><p>Traditional freelance platforms often have high fees and dispute issues. A blockchain version uses an <strong>Escrow Smart Contract</strong> to hold funds and release them only when specific milestones are met.</p><ul><li><strong>The Tech Stack:</strong> Solidity (Smart Contracts), Hardhat (Development Environment), React/Next.js (Frontend), and Ethers.js (Blockchain interaction).</li><li><strong>Key Feature:</strong> A “Dispute Resolution” mechanism where a DAO (Decentralized Autonomous Organization) or a set of jurors can vote on project completion.</li><li><strong>End-to-End Flow:</strong> Client posts job $\rightarrow$ Freelancer accepts $\rightarrow$ Client deposits funds in contract $\rightarrow$ Freelancer submits work $\rightarrow$ Contract releases payment.</li></ul><h3>2. Transparent Supply Chain for Luxury Goods</h3><p>Counterfeiting is a trillion-dollar problem. This app tracks a product’s journey from the manufacturer to the end consumer using <strong>NFTs as “Digital Twins.”</strong></p><ul><li><strong>The Tech Stack:</strong> Polygon (for low gas fees), IPFS/Filecoin (to store product images/metadata), and a Mobile App (React Native) for scanning.</li><li><strong>Key Feature:</strong> QR code integration. Scanning a physical product triggers a call to the blockchain to verify its entire ownership history and authenticity.</li><li><strong>End-to-End Flow:</strong> Manufacturer mints NFT $\rightarrow$ Logistical updates recorded as metadata $\rightarrow$ Consumer verifies “Provenance” before purchase.</li></ul><h3>3. Automated Insurance Protocol (Parametric Insurance)</h3><p>This removes the need for manual claims. For example, “Flight Delay Insurance” where the smart contract automatically refunds the user if a flight is delayed by more than 2 hours.</p><ul><li><strong>The Tech Stack:</strong> Chainlink Oracles (to fetch real-world flight data), Solidity, and a dashboard to view active policies.</li><li><strong>Key Feature:</strong> Use of <strong>Oracles</strong>. This demonstrates how blockchain interacts with external, non-blockchain data.</li><li><strong>End-to-End Flow:</strong> User buys policy $\rightarrow$ Smart contract queries flight API via Chainlink $\rightarrow$ If delay detected, contract executes transfer() of payout immediately.</li></ul><h3>4. Peer-to-Peer Energy Trading Grid</h3><p>With the rise of solar panels, neighbors can sell excess electricity to one another without a central utility company.</p><ul><li><strong>The Tech Stack:</strong> Energy Web Chain or an Ethereum Layer 2, integrated with IoT simulations.</li><li><strong>Key Feature:</strong> <strong>Micro-transactions</strong>. Handling many small, frequent payments efficiently.</li><li><strong>End-to-End Flow:</strong> Smart meter records excess energy $\rightarrow$ Listing created on a local marketplace $\rightarrow$ Neighbor’s contract “buys” energy $\rightarrow$ Settlement happens instantly on-chain.</li></ul><p>For any of these, I recommend starting with a <strong>Local Testnet</strong> (like Anvil or Ganache). Once your smart contracts are stable, deploy them to a <strong>Public Testnet</strong> (like Sepolia or Amoy) before even thinking about a Mainnet launch. This ensures you aren’t spending real capital during the debugging phase.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9926b0b4e894" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Agents Vs Agentic AI]]></title>
            <link>https://adosi12.medium.com/ai-agents-vs-agentic-ai-60b1e6d33564?source=rss-7b54740e75bb------2</link>
            <guid isPermaLink="false">https://medium.com/p/60b1e6d33564</guid>
            <category><![CDATA[agents]]></category>
            <category><![CDATA[chatbots]]></category>
            <category><![CDATA[2026]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Aditi Dosi]]></dc:creator>
            <pubDate>Fri, 09 Jan 2026 16:27:48 GMT</pubDate>
            <atom:updated>2026-01-09T16:27:48.973Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/878/1*TszNiJL72vK23OW8-ieevA.png" /></figure><p>While the terms are often used interchangeably, in the landscape of 2026, the industry has drawn a clear distinction between them.1 Think of <strong>AI Agents</strong> as the specialized workers and <strong>Agentic AI</strong> as the intelligent management system that empowers them.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/275/0*TE9QeNxY5mB4slge" /><figcaption>AI Agents Vs Agentic AI</figcaption></figure><h3>1. AI Agents: The “Worker Bees”</h3><p>An AI Agent is a specific software entity designed to perform tasks. It has access to “tools” (like an API, a database, or a web browser) and uses an LLM to decide which tool to use to fulfill a request.</p><ul><li><strong>Characteristics:</strong> Narrow scope, follows a specific “persona,” and is usually triggered by a human.</li><li><strong>Example:</strong> A <strong>Support Agent</strong> that can look up a customer’s order and process a refund when a user asks for it. It is excellent at “doing,” but it doesn’t decide to process refunds on its own.</li></ul><h3>2. Agentic AI: The “Orchestrator”</h3><p>Agentic AI describes an entire system (or an AI behavior) that demonstrates true <strong>agency</strong>.6 It doesn’t just follow instructions; it creates its own sub-tasks, plans for long-term goals, and monitors its own progress.</p><ul><li><strong>Characteristics:</strong> High autonomy, situational awareness, and the ability to coordinate multiple “AI Agents.”</li><li><strong>Example:</strong> An <strong>Agentic Supply Chain System</strong>. It doesn’t wait for a human to report a delay. It monitors weather patterns, sees a storm, predicts a logistics bottleneck, and independently directs various <em>Logistics Agents</em> to reroute shipments.</li></ul><h3>The 2026 Shift: From Tools to Ecosystems</h3><p>As we are seeing a shift away from standalone agents toward <strong>Agentic Orchestration</strong>.</p><p>Key trends include:</p><ul><li><strong>Standardized Protocols:</strong> New standards like the <strong>Agent-to-Agent (A2A) Protocol</strong> and Anthropic’s <strong>Model Context Protocol (MCP)</strong> allow different agents from different companies to “talk” to each other seamlessly.</li><li><strong>Agentic Workforces:</strong> Companies no longer buy one “AI Agent.” They build an “Agentic Framework” where dozens of specialized agents collaborate.</li><li><strong>Bounded Autonomy:</strong> The focus has moved to “guardrails.” We now design agentic systems with specific financial and ethical limits so they can act independently without needing a human to click “approve” on every tiny step.</li></ul><h3>Which one do you need?</h3><ul><li><strong>Use AI Agents</strong> if you have a repeatable, well-defined workflow that needs automation (e.g., “Summarize every email from this sender”).13</li><li><strong>Use Agentic AI</strong> if you have a high-level business objective that requires reasoning, constant monitoring, and cross-departmental action (e.g., “Reduce our customer churn rate by 10%”)</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=60b1e6d33564" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>