Using GitHub to Demonstrate Competence

For folks who use GitHub and are in early stage careers and hope to add GitHub as a value add to your profile - here is a note. If interviewing for an SDE role, GitHub projects that don't solve a problem and are just a coding exercise are not very helpful. This may sound perplexing but it needs to be said. : Hiring managers and tech leads (like me) look for problem-solvers. A repository full of practice exercises might show you can write code, but it doesn’t demonstrate that you can build impactful solutions. ► How to Make Your Projects Stand Out 1. Frame Them as Solutions: Instead of presenting your project as "just another app," position it as a business solution or a tool that solves a real-world problem. For example: - Instead of “Expense Tracker App,” say, “A tool for freelancers to manage and categorize expenses for tax season.” - Instead of “Weather App,” frame it as, “A weather app optimized for agricultural planning with location-based crop suggestions.” 2. Highlight the Problem It Solves: Every great project starts with a problem. Make it clear what problem you identified and how your project addresses it. - Example: “This tool was designed for small business owners who struggle with automating their daily sales tracking.” 3. Show Quantifiable Value: Numbers tell a story. Include metrics like: - How much time/money the solution saves. - How many users it could potentially impact. - Any test data or feedback you’ve collected. - Example: “This app reduced invoice processing time by 35% in a real-world test case.” 4. Document It Well: A project is only as good as its README. Include: - A brief description of the problem it solves. - Key features. - Instructions on how to run/test it. - Screenshots, GIFs, or a demo link to bring it to life. Having a GitHub full of clone apps or unfinished side projects sends the wrong signal. It doesn’t show creativity, ownership, or impact, it shows you can follow tutorials, and that’s not what companies hire for. Instead, invest your time into one or two high-impact projects that: - Solve real-world problems. - Show off your ability to understand user needs. - Demonstrate your thought process, design skills, and technical execution. So, take a step back, revisit your GitHub, and think: - Does this project solve a problem? - Can I explain its value to someone outside of tech? - Would I hire someone based on this work? If the answer isn’t “yes,” it’s time to rethink how you approach your projects. Remember: One impactful project > 100 clones. Focus on impact, not just output.

Only 24% of PM candidates have a GitHub. Every PM I've placed at OpenAI, Anthropic, and Meta AI in the last year had one. They didn't have 78K stars. They had three things: one project that solved a real problem, documentation that showed product thinking, and consistent activity. Shubham Saboo went from Dev Rel to Senior AI PM at Google because his GitHub marketed him. Google reached out to him. When companies come to you, the interview-to-offer rate jumps from 22% to 37%. But most PM GitHubs I review are anti-signals. Copy-pasted AI code with no tradeoffs section. Joke projects that prove nothing. Green square farms with 50 empty commits. I put together the web's first guide on PM GitHubs: • The 6 elements that made Shubham's GitHub get him hired • Step-by-step setup with Cursor and Claude Code • A starter kit repo you can fork today • 50 project ideas mapped to PM skills • The 6 mistakes I see candidates make Full guide for paid subscribers: https://lnkd.in/gzXDiXNE

Your GitHub profile is now your resume for high-paying AI roles. But most PMs don't know how to build one. They want roles at Google, OpenAI, Anthropic. But they have: • No public proof • No working prototypes • No technical surface area • No evidence they can ship AI systems Meanwhile Shubham Saboo built his GitHub like a product. In result, google reached out to him and he landed an AI PM role. Not from cold applying. Not from networking his way in. From literally his GitHub. So if you want serious AI roles, here’s the anatomy of a PM GitHub that actually works: 1. Welcome Message (Your Above-The-Fold Hook) Stop writing: “PM passionate about AI.” Use this instead: Scope (what you build) + System (how you think) + Signal (proof you execute) Example: "I build AI agents and evaluation systems for real workflows." That’s positioning. 2. Your Headline (3–4 Lines That Define You) Not another boring job description. Use: Role + Build Capability + Technical Depth + Outcome Example: AI Product Manager Build → prototype → eval → production thinking Hands-on with RAG, agents, cost modeling This reframes you from “PM” to AI operator. 3. The Star Repo (80% of Your Leverage) One repo can change your career. Most PMs build “notes on LLMs.” Wrong. Build something real like: • AI meeting copilot • RAG + eval pipeline • Multi-agent workflow • AI cost dashboard or a collection of resources. And structure your README like this: Problem → Architecture → Why this approach → Tradeoffs → Demo → What I’d improve If I can see how you think about production tradeoffs, you’re ahead of 90% of PMs. 4. What You Work On (Define Your Surface Area) 5–10 bullets max. Please don't just put, “AI, innovation, product strategy.” Instead: • Retrieval augmented generation systems • Agent orchestration logic • Prompt optimization & eval frameworks • Model comparison testing • AI cost modeling This signals depth (and actually learn these skills as well so you can walk the walk. See the end of this post) 5. Pinned Repos (Your Front Page) Pin 2–3 max. Only repos that show: Idea → System → Tradeoffs → Iteration. If it doesn’t demonstrate lifecycle ownership, don’t pin it. 6. Contribution Chart (Technical Credibility) You don’t need daily commits. You need meaningful ones. A healthy contribution chart signals: “I’m close to execution.” 7. Recent Contributions (Evidence Layer) - Add screenshots. - Add stats. - Document iterations. - Visual proof > vague claims. If you want to dive deep into building your Github like a pro, go read Aakash Gupta's awesome newsletter covering all this in detail: https://lnkd.in/gZCX6WQR If you want to master all the AI skills so you can become an AI PM, have wonderful AI projects on your Gihub, and future-proof your career, you'll love our #1 AI PM Certification: https://lnkd.in/gWQrdy-X

Why Your GitHub Is More Important Than Your CGPA - Low CGPA? - Not from a top-tier college?  - No formal CS degree? You’re not alone, and you’re not out of the game. The truth? Your GitHub can open more doors than your CGPA ever will. Here’s why GitHub > Grades ✅ It shows what you’ve actually built Not textbook answers or group assignments. Real code. Real projects. Real you. ✅ It tells your story Your commit history isn’t just code - it’s proof of your curiosity, consistency, and growth. ✅ It validates your skills No fluff. No buzzwords. Just cold, clean, beautiful code that hiring managers love to see. ✅ It’s your fingerprint Degrees can look the same - your GitHub is uniquely yours. That’s your edge. A recruiter doesn’t say: “Wow, an 8.9 CGPA.” They say: “Wait… you built and deployed a full-stack app by yourself? Let’s talk.” CGPA is just a number.  GitHub is proof. So, here’s what you should do: • Finish that side project. • Push that next commit. • Polish that README. • Document your process. You don’t need permission to be great - just the courage to build like you already are. Build loud. Build often. Build proud. #BuildInPublic #GitHubPortfolio #CodeOverCredentials #DevJourney #ProjectsOverCGPA #SelfTaughtDev #TechCareers #LearnByBuilding #CodingLife #SoftwareEngineer

Want your GitHub to stand out in interviews? Contribute where companies already hire from. Because recruiters don’t just check your GitHub activity, they check where you contribute. Building your own projects is great, it shows you can execute. But joining an open-source project shows you can collaborate, read someone else’s code, respond to feedback. That’s a different kind of signal. Here’s the simplified approach: 1️⃣ Pick a project you use or like. It could be something simple, a Python library, a web app, or a tool you’ve used before. Choosing a project you already understand makes it easier to start. 2️⃣ Look for an easy task. On GitHub, most projects tag beginner-friendly tasks as “good first issue” or “help wanted.” These are small fixes, correcting a typo, improving docs, or solving a minor bug. 3️⃣ Make a small change and submit it. Edit the file, save your change, and send it to the main project through a “Pull Request.” That’s basically saying, “Hey, I made this improvement. Can you add it?” Simple, isnt it? This week find one open-source repo, copy it, submit one change (code or documentation). Then next week screenshot your PR, add it to your LinkedIn/GitHub profile. Continue doing this, and see the difference. Check out comments for additional resources!

I’ve spent years observing how the best tools and strategies can help you stand out and land your dream cloud role.  Here’s what I learned: *** 𝗠𝗲𝗱𝗶𝘂𝗺: 𝗧𝗲𝗹𝗹 𝗬𝗼𝘂𝗿 𝗦𝘁𝗼𝗿𝘆 Documentation is your secret weapon. Use Medium to share your cloud projects, showcase your thought process, and let your work speak for itself.  Don’t just describe the end result, explain your 𝗦.𝗢.𝗔.𝗥.𝗙 process:   • 𝗦𝗰𝗲𝗻𝗮𝗿𝗶𝗼: What problem were you solving?   • 𝗢𝗯𝘀𝘁𝗮𝗰𝗹𝗲: What challenges did you face?   • 𝗔𝗰𝘁𝗶𝗼𝗻: What steps did you take to solve it?   • 𝗥𝗲𝘀𝘂𝗹𝘁: What did you achieve?   • 𝗙𝗮𝗶𝗹𝘂𝗿𝗲: What didn’t work, and how did you adapt?  Tip: Add screenshots and detailed steps to make it even more engaging.  *** 𝗚𝗶𝘁𝗛𝘂𝗯: 𝗖𝗼𝗺𝗺𝗶𝘁 𝘁𝗼 𝗘𝘅𝗰𝗲𝗹𝗹𝗲𝗻𝗰𝗲 Your GitHub profile is a testament to your growth and proof of your codebase.  𝗛𝗲𝗿𝗲’𝘀 𝗵𝗼𝘄 𝘁𝗼 𝘀𝘁𝗮𝗻𝗱 𝗼𝘂𝘁:   • Keep your commits consistent, regular updates show you’re improving.   • Add clear descriptions of your code and a READme file, what it does and why it matters.  Think of it as a portfolio that showcases both your technical skills and your problem-solving mindset.  Profiles with regular commits and well-documented READMEs are 5x more likely to catch a recruiter’s eye. 𝗦𝘁𝗿𝘂𝗴𝗴𝗹𝗶𝗻𝗴 𝘁𝗼 𝗰𝗿𝗲𝗮𝘁𝗲 𝗮 𝘀𝘁𝗮𝗻𝗱𝗼𝘂𝘁 𝗥𝗘𝗔𝗗𝗠𝗘? Use this GitHub README template to structure your project descriptions clearly and professionally.: https://lnkd.in/eWjmJqjC *** 𝗖𝗵𝗮𝘁𝗚𝗣𝗧: 𝗦𝗶𝗺𝗽𝗹𝗶𝗳𝘆 𝘁𝗵𝗲 𝗖𝗼𝗺𝗽𝗹𝗲𝘅 Explaining your projects doesn’t have to be intimidating. Tools like ChatGPT can help break down technical concepts into simple, digestible language.  𝗧𝗿𝘆 𝘁𝗵𝗶𝘀 𝗽𝗿𝗼𝗺𝗽𝘁:   "𝘙𝘦𝘸𝘳𝘪𝘵𝘦 𝘵𝘩𝘪𝘴 𝘵𝘰 𝘦𝘹𝘱𝘭𝘢𝘪𝘯 𝘢𝘵 𝘢 5𝘵𝘩-𝘨𝘳𝘢𝘥𝘦 𝘭𝘦𝘷𝘦𝘭.”  Your goal isn’t to sound like a genius, it’s to ensure anyone following along understands the value of your work.  *** 𝗔𝘂𝘁𝗵𝗼𝗿𝗲𝗱𝗨𝗽: 𝗦𝗵𝗮𝗿𝗲 𝘄𝗶𝘁𝗵 𝗜𝗺𝗽𝗮𝗰𝘁  When it’s time to share your project on LinkedIn, storytelling matters. Craft posts that answer the reader’s unspoken question: “𝗪𝗵𝗮𝘁’𝘀 𝗶𝗻 𝗶𝘁 𝗳𝗼𝗿 𝗺𝗲?”  For example:   • 𝗜𝗻𝘀𝘁𝗲𝗮𝗱 𝗼𝗳: “Check out my new project!”   • 𝗧𝗿𝘆: “This is how I deployed AWS Infrastructure in less than 10 seconds.”  Build a story around how your project helped you solve 𝗫, 𝗬, 𝗼𝗿 𝗭. You’ll resonate more with your audience and inspire them to engage.  *** Landing your next cloud role isn’t just about showing what you’ve built, you have to be able to articulate how you solved problems and how you improved on solving that problem.  Don't forget, its all about building your 𝗕𝗥𝗔𝗡𝗗! Using tools that can make it easier for you to stand out. What's your tech stack that you use to build your brand on LinkedIn.

Your GitHub portfolio is not a novel. Stop writing it like one.   I recently had a coaching session where someone showed me their GitHub project. Super technical, lots of great work. But then I opened the repo and boom… wall of text. One long README. No visuals. No context. Just line after line of dense explanations.   I told him: “This is solid, but not visual. You need to help people see what you’ve built, not just read about it.”   Because let’s be honest: nobody reads every word on your GitHub unless they’re already sold on you. Most hiring managers? They scroll. They skim the headlines. They want a quick visual confirmation that you know your stuff.   So here’s a simple rule: 👉 If your GitHub reads like a textbook, you’ll lose attention. 👉 If it looks like a dashboard - clean, visual, and easy to navigate - you’re on the right track.   Add screenshots of dashboards. Drop in architecture diagrams. Link to dbt docs. Show a dataset preview. Visual cues help people connect the dots fast.   Remember: your project might be technically perfect. But if no one can understand it in 30 seconds, it’s not working for you.   Make it scrollable. Make it visual. Make it easy. * You’re unsure how to make your GitHub stand out? 👉 That’s exactly what we work on in my Data Engineering Coaching program. We help you build solid projects and present them in a way that gets you noticed. Check out the program and apply via the link in the comments! *

My client passed 8 out of his next 10 technical assessments in just 4 weeks of working together They went from failing every technical assessment, hating and blaming the system… But the truth is: You don't rise to the occasion. You fall to the level of your preparation. So I taught him what I teach all my clients: Don’t cram for interviews, train like an athlete, and practice like a professional. Here’s how I help clients prep for interviews without burning out or waiting until an assessment shows up in their inbox. We build coding prep like a habit stack. Each layer trains a real-world interview skill. 𝟭) 𝗟𝗲𝘃𝗲𝗹 𝟭: 𝗗𝗮𝗶𝗹𝘆 𝗰𝗼𝗱𝗶𝗻𝗴 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗲 (𝟯𝟬 𝗺𝗶𝗻𝘀) Daily coding practice on Platforms WHY? To start interview prep for the assessment and live coding rounds 𝟮) 𝗟𝗲𝘃𝗲𝗹 𝟮: 𝗣𝗿𝗮𝗰𝘁𝗶𝗰𝗲 + 𝗽𝘂𝘀𝗵 𝘁𝗼 𝗚𝗶𝘁𝗛𝘂𝗯 Daily coding practice on Platforms + Git, committing progress\ WHY? To retain more information and be able to reflect on your progress 𝟯) 𝗟𝗲𝘃𝗲𝗹 𝟯: 𝗔𝗱𝗱 𝗰𝗹𝗲𝗮𝗿 𝗰𝗼𝗺𝗺𝗶𝘁 𝗺𝗲𝘀𝘀𝗮𝗴𝗲𝘀 Daily coding practice on Platforms + Git, committing progress with good commit messages WHY? To practice clearly communicating to your team with git 𝟰) 𝗟𝗲𝘃𝗲𝗹 𝟰: 𝗧𝗿𝗮𝗰𝗸 𝘁𝗶𝗺𝗲 𝗶𝗻 𝗰𝗼𝗺𝗺𝗶𝘁 𝗺𝗲𝘀𝘀𝗮𝗴𝗲𝘀 Daily coding practice on Platforms + Git, committing progress with good commit messages + Adding time, then adding how long it took in your commit message WHY? To practice like the interview (assessment or live coding) with a little more nervousness, so you can crush the interviews 𝟱) 𝗟𝗲𝘃𝗲𝗹 𝟱: 𝗧𝗮𝗹𝗸 𝗼𝘂𝘁 𝗹𝗼𝘂𝗱 𝘄𝗵𝗶𝗹𝗲 𝘀𝗼𝗹𝘃𝗶𝗻𝗴 Daily coding practice on Platforms + Git committing progress with good commit messages + Adding time, then adding how long it took in your commit message + Talking aloud WHY? To practice like the interview (live coding, case study, system design) by communicating your actions and reasoning, so your practice is more like the real thing, and you can crush the interviews 𝟲)𝗟𝗲𝘃𝗲𝗹 𝟲: 𝗦𝗵𝗮𝗿𝗲 𝘄𝗵𝗮𝘁 𝘆𝗼𝘂 𝗹𝗲𝗮𝗿𝗻𝗲𝗱 (𝗮𝘀 𝗰𝗼𝗻𝘁𝗲𝗻𝘁) Daily coding practice on Platforms + Git committing progress with good commit messages + Adding time, then adding how long it took in your commit message + Talking aloud + Creating a mini content WHY? To start building the habit of creating content from your learnings. The easiest way to create content is to document your life This method has helped my clients pass 80 %+ of coding rounds and land roles in data, analytics, and tech. Take a look at what stage you are on, then, when you are consistent, you can move to the next one. Let’s build the habit, not the panic. ♻️ Repost if you found this helpful

This post is for all #junior #AI engineers. Working on sample projects and publishing them to GitHub is a great way to gain experience and increase your visibility. But be very mindful of what you share publicly. The code you post online reflects your software engineering and AI skills and can be used against you. I've come across many projects with good ideas, but as soon as I look at the code, I spot fundamental software engineering mistakes. One common example: exposed API keys in the repository. If I were a recruiter, seeing a leaked API key would immediately raise red flags. If someone can't safeguard something as basic as an API token, how could they be trusted with more sensitive or critical information? A simple fix? Use the pydantic-settings package. It provides a BaseSettings class that helps you manage secrets securely through environment variables. No more accidental leaks for now. We'll discuss the other mistakes later.