<?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 Thomas To on Medium]]></title>
        <description><![CDATA[Stories by Thomas To on Medium]]></description>
        <link>https://medium.com/@thomas.to.bcheme?source=rss-07bc157dafb8------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*q77v2IBSEDVXywCP</url>
            <title>Stories by Thomas To on Medium</title>
            <link>https://medium.com/@thomas.to.bcheme?source=rss-07bc157dafb8------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 16:44:15 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@thomas.to.bcheme/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[Building a Public Portfolio as a Fullstack AI Engineer]]></title>
            <link>https://medium.com/@thomas.to.bcheme/building-a-public-portfolio-as-a-fullstack-ai-engineer-c8957f95a9e7?source=rss-07bc157dafb8------2</link>
            <guid isPermaLink="false">https://medium.com/p/c8957f95a9e7</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[open-source]]></category>
            <category><![CDATA[data-engineering]]></category>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[personal-branding]]></category>
            <dc:creator><![CDATA[Thomas To]]></dc:creator>
            <pubDate>Mon, 20 Jul 2026 03:35:39 GMT</pubDate>
            <atom:updated>2026-07-20T03:35:39.352Z</atom:updated>
            <content:encoded><![CDATA[<p>A practical stack for turning scattered side projects into a public portfolio other engineers find easily.</p><h3>Abstract</h3><p>The article argues that most engineers do good technical work but fail to make it discoverable or credible to people outside their immediate circle. It lays out a five-part stack, presented in the order the author recommends building it, for closing that gap.</p><p>1. LinkedIn — Framed as a searchable index rather than a resume. Recommends a specific headline, a plain-language About section,<br> matching skills, and regular engagement with others’ posts (not just self-broadcasting), citing LinkedIn’s own stats on profile<br> completeness and photos driving views.</p><p>2. A free-tier portfolio stack — GitHub, Vercel, and Neon combine to host a live, working project at no cost. The piece stresses that<br> a portfolio needs more than a repo link: a deployed app, a README explaining tradeoffs, a visible data model or API contract, a<br> working CI/CD pipeline, and honest notes on what broke during development.</p><p>3. Open source contribution — Counters the assumption that contributions must be significant to matter. Points to GitHub’s /contribute<br> URL suffix for beginner-tagged issues, and highlights Manning’s MEAP program (early access to in-progress technical books with<br> companion code repos) as a low-intimidation entry point for a first pull request.</p><p>4. Developer communities — Discord/Slack communities built around a stack already have the target audience. The advice is to join a few relevant ones, answer questions before asking or promoting, and use a consistent handle to build recognition over time, rather than treating communities as another broadcast channel.</p><p>5. Earned media — Positioned as the compounding, final layer rather than a starting point. PR agencies (the article cites Pathos Communications’ pay-on-coverage model) need a real body of public work to pitch in the first place, so this step only works once the other four are in place.</p><p>You finish a project, push it to GitHub, and open a tab to add it to your portfolio. Three sentences in, you stop. The work is real, but nothing about how it lives online makes this obvious to anyone who was not in the room when you built it.</p><p>This gap between doing the work and pointing to it is common among engineers. Most of us are trained to write code, not to document proof of it in a way other people find easily. When a collaborator, a conference organizer, or another engineer in your field looks you up, what they find is usually a half-finished LinkedIn profile, a GitHub with fifteen repositories named project, and no clear thread connecting any of it.</p><p>None of the individual pieces are hard. LinkedIn, a public portfolio, open source, developer communities, and public relations each solve one part of the same problem: making real work visible to people who were not there when you built it. What follows is a walkthrough of each piece, in the order I found most useful to build them.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/877/1*bkRvvPsp9oSUHxWv0qw6oQ.png" /><figcaption>Five separate proofs of work, LinkedIn, a portfolio, open source, communities, and earned media, only matter if they point somewhere. Together they build one credibility record other engineers can find without ever having watched you build it.</figcaption></figure><h3><strong>LinkedIn as your searchable professional record</strong></h3><p>LinkedIn works less like a resume and more like a search index. Recruiters, collaborators, and other engineers do not read your profile top to bottom. They scan it for signals matching what they want, then move on when those signals are missing.</p><p>The data backs this up. LinkedIn’s own numbers show a complete profile receives significantly more views and messages than an incomplete one, and profiles with a photo receive up to 21 times more views than profiles without one [1]. A strong, specific headline alone accounts for a measurable share of the difference.</p><h4><strong>Profile details worth prioritizing</strong></h4><p>A handful of details do most of the work:</p><p>- A headline naming the stack or domain you work in, not a generic title like Software Engineer</p><p>- An About section with two or three concrete outcomes, written in plain language instead of buzzwords</p><p>- A skills section matching the technologies visible in your project work</p><p>- Regular engagement with other people’s posts, not only publishing your own</p><p>- A consistent posting cadence, even if it is only once every week or two. This can be automated using the linkedin API to post text and github actions. See my open source repo to get started, <a href="https://github.com/thomas-to-bcheme/linkedin-content-cicd">https://github.com/thomas-to-bcheme/linkedin-content-cicd</a></p><h3><strong>A portfolio built on the free tier</strong></h3><p>LinkedIn tells people you exist. A portfolio shows what you build. The gap between the two is where most engineers lose credibility, because a list of technologies on a profile is not evidence of anything.</p><p>The infrastructure to close this gap costs nothing at the scale most individual engineers operate at. GitHub holds the code and the commit history. Vercel serves the frontend and handles blob storage for static assets. Neon provides a serverless Postgres database which scales to zero when nobody is using it [2]. Together, these three services let you deploy a working fullstack or agentic application using Gemini without paying for a server you are not using yet.</p><p>The free tier limits are generous, and they rarely become the constraint. Neon’s free plan includes up to 100 projects and roughly half a gigabyte of storage per branch, and Vercel’s Hobby plan includes 100GB of bandwidth and 100 hours of function execution a month [2]. A single portfolio project, fits comfortably inside those numbers.</p><h4><strong>What belongs in the portfolio</strong></h4><p>A repository link by itself proves little. A stronger case for your work includes:</p><p>- A live, deployed version of the project, not only a link to source code</p><p>- A README explaining the decisions you made and the tradeoffs you weighed, not only how to run it</p><p>- A visible data model or API integration</p><p>- At least one project running through a CI and CD pipeline end to end</p><p>- A short note on what broke during development, how you fixed it, and lessons learned</p><p>This last point is the one engineers skip most often, and it is usually the most convincing. Many engineers ship a project which works. Writing honestly about what did not work the first time signals real experience with the system, not only familiarity with a tutorial.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/912/1*gFrs6hAiMyUC6wluR1EXmA.png" /><figcaption>Deploy pipeline for a real project: push to GitHub, run through CI/CD, ship to Vercel, which serves the app, calls Gemini for production agentic AI, and queries Neon for data.</figcaption></figure><h3><strong>Contributing to open source without the intimidation</strong></h3><p>Open source is the part of this list most engineers want to do and avoid starting. The instinct is to assume you need to fix something significant on a project with tens of thousands of stars before anyone takes a pull request seriously. This assumption keeps most people from ever opening one.</p><p>Most projects make the entry point easier than it looks. Adding `/contribute` to the end of a GitHub repository’s URL surfaces issues maintainers have specifically tagged as approachable for newcomers [3]. Starting there, on documentation fixes or small bugs, teaches you how a project’s maintainers review code and communicate before you attempt anything larger.</p><h4><strong>Manning’s MEAP</strong></h4><p>Manning Publications runs a program called MEAP, the Manning Early Access Program, where readers get access to a technical book chapter by chapter while the author is still writing it [4]. Many of these books ship with a companion code repository on GitHub, evolving alongside the manuscript.</p><p>This combination creates an unusual opening. You read a chapter, work through the accompanying code, find a bug or an outdated dependency, and submit a pull request directly against the author’s repository while the book is still in progress. The barrier to entry is lower than most open source projects because you are already deep in the material by the time you find something worth fixing, and the author is actively watching the repository.</p><p>A short list of ways to start: <a href="https://www.manning.com/meap-catalog">https://www.manning.com/meap-catalog</a></p><p>- Sign up to be a manning reviewer <a href="https://www.manning.com/manuscript-reviews">https://www.manning.com/manuscript-reviews</a> and in support, get complimentary copies of MEAP books.</p><p>- Choose a MEAP book in a topic you are learning and read its code repository alongside the chapters</p><h3><strong>Finding your place in existing developer communities</strong></h3><p>Discord and Slack communities built around specific frameworks, languages, or engineering disciplines already have the audience you are trying to reach. Joining one is free. Being useful in one is what builds a reputation.</p><p>The mistake is treating these spaces as another channel to broadcast into. A community notices the people who answer questions, not the people who only post their own project once a month. Spending the first few weeks in a new community helping other people with their problems, before mentioning any of your own work, changes how your later posts land.</p><p>A few habits make this easier to sustain:</p><p>- Join two or three communities mapping directly to your stack instead of a dozen general ones</p><p>- Answer questions you know the answer to before asking your own</p><p>- Share your own project only after you have helped a handful of other people first</p><p>- Use a consistent handle and avatar across platforms so people recognize you over time</p><p>This is slower than posting a project announcement and hoping it gets traction. It is also the part of the process producing the kind of relationships lasting past a single post.</p><p>Here are a few notable communities I am apart of:</p><ol><li><a href="https://datatalks.club/slack.html">https://datatalks.club/slack.html</a></li><li><a href="https://discord.gg/FuSYewDZud">https://discord.gg/FuSYewDZud</a> (Chip Huyen MLOps)</li><li><a href="https://huggingface.co/discord-community">https://huggingface.co/discord-community</a></li></ol><h4><strong>Earned media as the compounding layer</strong></h4><p>Everything up to this point is work you do directly. Earned media is different. It is a third party, usually a journalist or outlet, deciding your work is worth covering, and it is the part of this process most engineers never consider because it seems disconnected from writing code.</p><p>Public relations agencies exist specifically to make this connection. Some, like Pathos Communications, have restructured the traditional PR model entirely, billing clients only after they secure coverage in a publication the client values, instead of charging a retainer upfront regardless of outcome [5]. This model matters for individual engineers because it removes the financial risk of trying.</p><p>What makes an agency’s job possible is exactly what the first four pillars produce. A PR agency struggles to pitch a story about someone with no public work to point to. A live project with a real architecture, a pull request history on projects people recognize, and a pattern of community engagement gives a pitch something concrete to describe. A feature in an outlet like International Business Times or a comparable trade publication becomes a data point which the LinkedIn profile, the portfolio, and the open source history all link back to.</p><p>This is why it belongs last. Earned media does not create credibility from nothing. It amplifies credibility already built, and it compounds fastest once the other four pieces are in place.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6k-wJdRC__UYZ_TlHeF_7g.png" /></figure><p><a href="https://www.ibtimes.com/thomas-how-ai-enabled-biology-advancing-drug-discovery-scientific-innovation-3804756">https://www.ibtimes.com/thomas-how-ai-enabled-biology-advancing-drug-discovery-scientific-innovation-3804756</a></p><p>Otherwise some other ways to “put yourself out there” is posting on social media and creating content. This could be your own thoughts, experiences as a industry professional, creating educational content, or explaining your projects in video format to build in public without having to pay.</p><p>There are many social media platforms and resources such as notes on the interview process, data structures and algorithms, system design, and general career advice. This includes but is not limited to YouTube, Instagram, X (twitter), or Reddit, in addition to the previously mentioned linkedin, discord, slack, and other communities.</p><p>Taking my own advice, medium is a more elaborate platform than linkedin and leveraging my own portfolio’s architecture to support credibility for opportunities I would not have gotten otherwise from interviewing. My <a href="https://www.youtube.com/@thomas-to-bcheme">https://www.youtube.com/@thomas-to-bcheme</a> and other content generating platforms have been on the back burner as I continue to maximize the free end to end architecture to deploy production agents for personable use &amp; showcase how I am scaling agents to support applying to jobs in this new AI era.</p><h3><strong>Key Takeaways</strong></h3><p>By building a foundation on linkedin for passive searchability, recruiters and connections come to you. The portfolio serves as credibility but can also serve as your own personable fullstack platform to continue supporting the job search with linkedin posting, your own job board and agentic continous delivery of tailored resumes, vectorless RAG chatbot, and practice integrating external open source to maintain for your needs (e.g excalidraw)</p><p>- A complete, specific LinkedIn profile receives measurably more views and messages than a generic one, and engagement matters as much as the profile itself.</p><p>- A free hosting stack of GitHub, Vercel, Gemini, and Neon is enough to deploy and demonstrate real fullstack or agentic work without any infrastructure cost.</p><p>- Open source contribution has a lower barrier than it appears, and programs like Manning’s MEAP offer an approachable entry point tied directly to material you are already learning.</p><p>- Developer communities on Discord and Slack reward consistent, reciprocal engagement over occasional self-promotion.</p><p>- Earned media through a PR agency works best as the final step, since it depends on having real, visible work to pitch in the first place.</p><p>Building each of these pieces takes time, and none of them require waiting for permission to start. If you found this useful, follow for more on AI/ML and Software Engineering.</p><p>References:</p><p>[1] LinkedIn Profile Statistics and Engagement Data 2026 — <a href="https://www.cognism.com/blog/linkedin-statistics">https://www.cognism.com/blog/linkedin-statistics</a></p><p>[2] Deploy Anything for $0: The Vercel + Neon + GitHub Actions Stack — <a href="https://www.birjob.com/blog/deploy-for-zero-dollars-vercel-neon-github-actions">https://www.birjob.com/blog/deploy-for-zero-dollars-vercel-neon-github-actions</a></p><p>[3] How to Contribute to Open Source — Open Source Guides — <a href="https://opensource.guide/how-to-contribute/">https://opensource.guide/how-to-contribute/</a></p><p>[4] MEAP Program — Manning Publications — <a href="https://www.manning.com/meap-program">https://www.manning.com/meap-program</a></p><p>[5] Pathos Communications <a href="https://payonresultspr.com/">https://payonresultspr.com/</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c8957f95a9e7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Software and Data Interview Preparation]]></title>
            <link>https://medium.com/@thomas.to.bcheme/software-and-data-interview-preparation-664b8bc04025?source=rss-07bc157dafb8------2</link>
            <guid isPermaLink="false">https://medium.com/p/664b8bc04025</guid>
            <category><![CDATA[career-development]]></category>
            <category><![CDATA[data-engineering]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[job-search]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Thomas To]]></dc:creator>
            <pubDate>Tue, 26 May 2026 14:01:02 GMT</pubDate>
            <atom:updated>2026-05-26T14:01:02.576Z</atom:updated>
            <content:encoded><![CDATA[<h3>Abstract</h3><p>The job market converts fewer than 1% of applications into offers. This article maps the full technical interview preparation process for Data and Software Engineering related roles, starting before the first recruiter call. It covers optimizing your resume and LinkedIn profile for ATS screening, conducting compensation research using Levels.fyi, and preparing for common recruiter questions including visa status, start date, and right-to-represent agreements. From there, it breaks down the three technical evaluation formats: fundamentals, system design, and behavioral. Each section outlines what evaluators assess and where to focus preparation time. The article references open source tools for tailoring application materials, an exploratory analysis of 2025 job market salary distributions, and learning resources available through O’Reilly and Packt via public library partnerships. This is the first in a series sharing studying notes, resource reviews, and lessons learned from each stage.</p><h3>In this article:</h3><h4>Before the First Technical Screen</h4><ul><li>Updating Resume and LinkedIn</li><li>Market Research on Compensation</li><li>The Recruiter Screen</li></ul><h4>The Three Interview Formats</h4><ul><li>Technical Fundamentals</li><li>System Design</li><li>Behavioral</li></ul><h4>Learning Resources</h4><ul><li>O’Reilly and Packt</li></ul><h4>Key Takeaways</h4><p>As active job seekers to Data and/or Software engineering roles, employers assess three distinct skill sets across three separate formats: technical fundamentals, system design reasoning, and behavioral fit. Each format operates by different rules and rewards different preparation. However, this assumes candidates have the opportunity to interview.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/263/1*pNTvC_ZIj9HQ28aTJJHcDw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/418/1*H7QU8PZzQdG4EOVMCtXv9A.png" /><figcaption><strong>Left </strong>(The &lt;1% Reality): The standard application funnel converts at under 1%. From 5,000 submissions, only 5% pass ATS screening, 1% reach recruiter review, and fewer than 0.4% get shortlisted. <strong>Right</strong> (Hybrid Reach Matrix): Four quadrants map visibility against outreach direction. Cold applying is low-visibility outbound. Active networking and referrals move you to high-visibility outbound. Brand optimization through LinkedIn content shifts inbound traffic from passive waiting to active recruiter discovery.</figcaption></figure><p>This article provides an overview for the subsequent medium articles documenting my experiences through the interview process. Through this process, the tools and approach outlined here have helped me consistently reach the shortlisted stage and focus my energy on interviewing preparation. This series documents the full process: each following article goes deeper on a specific area, sharing studying notes, resource reviews, and first-hand observations from O’Reilly available through my local public library partnership.</p><p>Follow along as I share lessons learned from actively interviewing.</p><h3>Before the First Screen</h3><h4>Updating Resume and LinkedIn</h4><p>Your resume and LinkedIn profile are filtered before a recruiter reads them. Applicant tracking systems match keywords from the job description against your submitted materials. Roles in data and software engineering list specific technologies, frameworks, and methodologies. Matching the language in the posting improves your visibility in recruiter searches and improves your chances of passing automated screening.</p><p>Tailoring materials to each role is more effective than sending a single generic version. I built open source tools to support this process:</p><ul><li><strong>Resume tailoring:</strong> generates a version of your resume aligned to a specific job description’s requirements and vocabulary</li><li><strong>Cover letter generation</strong>: writes a focused letter using language from the posting and your background</li><li><strong>Job description response</strong>: structures answers to prompted questions</li></ul><p><a href="https://github.com/thomas-to-bcheme/resume">https://github.com/thomas-to-bcheme/resume</a></p><p>A strong LinkedIn profile complements a tailored resume. Recruiters search LinkedIn using the same keywords they pulled from the job description. Your headline, About section, and experience descriptions need to reflect the language used in the roles you are targeting.</p><h4>Market Research on Compensation</h4><p>Entering a negotiation without data puts you at a disadvantage. Before applying, research competitive compensation for the specific role, level, and location using Levels.fyi <a href="https://www.levels.fyi/">[1]</a>. The platform aggregates self-reported compensation from verified employees at technology companies, broken down by base salary, equity, and bonus components, sorted by company, role, and level.</p><p>Having a specific range ready before the first recruiter call means entering that conversation with a potential role you would actually accept. Otherwise, candidates can effectively go through the interviewing process for practice.</p><p>Alternatively, I have prepared an analysis of the 2025 job market posting and plan to provide ongoing support for year over year analysis of the market <a href="https://www.linkedin.com/feed/update/urn:li:activity:7419433022703300609/">[2]</a>.</p><h4>Recruiter Frequently Asked Questions</h4><p>Questions to have answers ready for before the first call:</p><ul><li><strong>Visa status:</strong> Employers ask this to confirm work authorization. Know your status and be precise about any restrictions or timelines.</li><li><strong>Available start date: </strong>Two weeks from the date of a signed offer is the standard professional expectation. This signals respect for your current commitments and fits most employers’ onboarding timelines.</li><li><strong>Compensation expectations</strong>: Have a specific range ready based on your market research. Recruiters use this to confirm budget alignment early and avoid advancing candidates whose expectations fall outside the role’s budget.</li><li><strong>Role scope alignment</strong>: Be ready to describe your current scope and what you are targeting in your next role.</li><li><strong>Right to Represent: </strong>For roles sourced through third-party contracting firms, expect a request to sign a right-to-represent agreement before your profile gets submitted. This document authorizes the firm to present your candidacy to their client. Read it before signing. These agreements typically restrict you from applying directly to the same company for a defined period, which matters if you would prefer to pursue a direct-hire path with that employer.</li></ul><h3>The Three Interview Formats</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/680/1*n0VQ8BOK4XZsXrESbfT6hQ.png" /></figure><p>Organizations assess candidates across three distinct formats. Each operates by different rules and rewards different preparation. Knowing what each format evaluates before preparation begins changes how you invest your study time.</p><h4>Technical Fundamentals</h4><p>Technical fundamentals rounds evaluate your ability to reason about code under time constraints. Questions draw from core computer science concepts: algorithms, data structures, complexity analysis, and problem decomposition. The format varies by company: shared code editors with algorithm problems, take-home projects, or live pair programming where the evaluator observes your process directly.</p><p>What the format tests:</p><ul><li>Algorithm implementation and time/space complexity analysis</li><li>Data structure selection and its effect on runtime performance</li><li>Problem decomposition: breaking ambiguous problems into solvable steps</li><li>Communication while solving: narrating decisions and tradeoffs out loud</li><li>Language-specific patterns from the technologies listed in the role’s stack</li></ul><p>The job description signals what to prioritize. A backend role referencing distributed systems and message queues weights different concepts than a frontend role listing rendering performance and state management patterns.</p><h4>System Design</h4><p>System design rounds assess architectural reasoning, scalability thinking, and tradeoff analysis. There is rarely one correct answer. The evaluator watches how you frame the problem, gather requirements, identify constraints, and justify each decision you make.</p><p>Common areas covered:</p><p>- API design: request/response contracts, versioning strategies, error handling</p><p>- Data modeling: storage selection, schema design, query pattern optimization</p><p>- Scaling approaches: horizontal scaling, caching strategies, load distribution</p><p>- Consistency and availability tradeoffs in distributed systems</p><p>- Failure modes, recovery patterns, and graceful degradation</p><p>System design preparation rewards breadth of exposure to design a complete end-to-end high level design. Depending on the seniority of role, it expected to have a second pass of the high-level design to have in-depth discussions in any single area. Reading engineering blogs on how companies scaled, understanding commonly used API documentation from real providers, and studying distributed systems fundamentals all contribute to the pattern library these rounds require.</p><h4>Behavioral</h4><p>Behavioral rounds evaluate communication, collaboration patterns, and organizational fit. The format is not less rigorous than technical rounds. It requires the same deliberate preparation applied to your work history instead of computer science material.</p><p>Common areas covered:</p><ul><li>Collaboration under ambiguity or competing priorities</li><li>Handling disagreement with colleagues or managers productively</li><li>Project ownership: managing scope, dependencies, and delivery under constraints</li><li>Learning from setbacks and iterating on processes</li><li>Alignment between your working style and the team’s operating norms</li></ul><p>Effective preparation means organizing past experiences into clear, specific narratives before the conversation, with measurable outcomes where your work produced them. The employer-employee relationship begins at this stage: evaluators assess whether collaboration would be productive and whether the role matches where you want to grow.</p><h3>Learning Resources</h3><p>O’Reilly Learning and Packt are available through many public library systems at no cost. Both platforms cover algorithms, data structures, system design, distributed systems, and applied engineering topics at the depth needed for technical preparation.</p><p>For algorithms and data structures, materials combining visual explanations with implementation examples build better intuition than text-only references. The goal is pattern recognition: seeing a problem’s structure and matching it to the right approach, not memorizing solutions.</p><p>For system design, two foundational resources available through O’Reilly: “Designing Data-Intensive Applications” by Martin Kleppmann <a href="https://www.oreilly.com/library/view/designing-data-intensive-applications/9781491903063/">[3]</a> and “System Design Interview” by Alex Xu <a href="https://www.amazon.com/System-Design-Interview-insiders-Second/dp/B08CMF2CQF">[4]</a>. Kleppmann covers storage engines, replication, and distributed systems fundamentals at depth. Xu covers the interview format itself with annotated worked examples. Both reward a full read rather than selective reference.</p><p>For behavioral preparation, structured communication frameworks help organize past experiences into coherent narratives before high-stakes conversations.</p><h3>Key Takeaways</h3><ul><li>Getting interviews requires optimized application materials before preparation for the interview itself matters</li><li>Tailor your resume and LinkedIn vocabulary to each job description to improve ATS visibility and recruiter response rates</li><li>Research compensation using Levels.fyi before the first recruiter call. A specific range gives you a clear floor in that conversation</li><li>Prepare answers before the recruiter screen: visa status, available start date (two weeks from offer), compensation expectations, and right-to-represent awareness for roles sourced through third-party firms</li><li>Technical interviews assess three distinct skill sets: fundamentals, system design, and behavioral. Each requires separate, focused preparation</li></ul><p>If you found this useful, follow for more on AI/ML and Software Engineering.</p><p>References:</p><p>[1] Levels.fyi — Compensation data for technology roles — <a href="https://www.levels.fyi/">https://www.levels.fyi</a></p><p>[2] Job Market Exploratory Data Analysis 2025 — <a href="https://www.linkedin.com/feed/update/urn:li:activity:7419433022703300609/">https://www.linkedin.com/feed/update/urn:li:activity:7419433022703300609/</a></p><p>[3] Martin Kleppmann, “Designing Data-Intensive Applications” — <a href="https://www.oreilly.com/library/view/designing-data-intensive-applications/9781491903063/">https://www.oreilly.com/library/view/designing-data-intensive-applications/9781491903063/</a></p><p>[4] Alex Xu, “System Design Interview — An Insider’s Guide” — <a href="https://www.amazon.com/System-Design-Interview-insiders-Second/dp/B08CMF2CQF">https://www.amazon.com/System-Design-Interview-insiders-Second/dp/B08CMF2CQF</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=664b8bc04025" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>