Marco Chen

Hot take: Clean code is KILLING your startup. I just reviewed 30 YC startup codebases. The most successful ones all have one thing in common: absolute garbage code. We're talking: - Functions with 200+ lines - Variables named temp, data2, finalFinal - Copy-pasted blocks everywhere - Zero tests - Comments that say "idk i'll fix this later lol" Meanwhile, the failed startups? They have beautiful code, perfect abstraction layers, and 100% test coverage. And they're all dead products. This is what's happening in 2025: The vibe-coders are winning. They're using Cursor or Claude Code to ship features in hours, not days. And they're using Willow (https://lnkd.in/gf4C4JjP) to literally dictate what they want: "Make a user dashboard that shows their last 10 orders with a delete button." The use of voice dictation to prompt Cursor makes the process 5x faster. AI then spits out working code. Is it elegant? No. Does it work? Yes. Traditional devs spend 3 days refactoring. Vibe-coders spend 3 days shipping 5 new features. Here's the brutal math: - Clean code: 1 feature/week - Vibe code: 5 features/week Guess who finds product-market fit first? Code quality literally does not matter until you hit 10 mil ARR. Before that, just ship and vibe your way to revenue, then refactor where you can afford a team. What do you think? Am I advocating for technical debt or just being realistic about what actually kills startups?

78

19 Comments

US founders hiring Flutter in US time zones: here’s a 20-minute screen that beats a week of interviews. Ask for a production story. Prompt: “Walk me through a production issue you personally drove to resolution. What broke. What you did first. What you shipped. What you changed after.” What you’re listening for: They speak in “I”, not “we.” They make decisions, not escalations. They can name the follow-up change, not just the fix. That’s senior in a startup. Years of Flutter are just a proxy.

16

7 Comments

Aspiring Frontend Engineers: A to-do app with a React tutorial isn't a portfolio. Hiring engineers reviewing your GitHub can tell. Trust me. Recruiters spend 11 seconds on your GitHub. In those 11 seconds, your to-do app looks identical to 4,000 other to-do apps built from the same React tutorial. Same component structure. Same state management. Same deployment. Same "clone and run npm start" README. You didn't build an app. You just followed a tutorial. In 2026, companies are shipping AI-powered interfaces, real-time collaboration tools, and apps that work offline on spotty mobile connections. Your portfolio is still stuck in the React Hooks launch era. Proof of Work Week Day 8: Frontend Engineering 5 frontend projects that act as proof: 1. Full Design System Component Library Build a reusable component library from scratch. Buttons, inputs, modals, dropdowns, toasts, all with consistent theming, accessibility baked in, and Storybook documentation. Make it so good another developer could install it and ship with it. Stack: React or Vue + Storybook + Tailwind or CSS Modules. 2. Performance-Optimized Landing Page Take any slow website (use Lighthouse to find one), rebuild it from scratch. Hit 95+ on all Lighthouse scores. Document every optimization decision you made and why. Show the before and after. Stack: Next.js or Astro + lazy loading + image optimization + Core Web Vitals tracking. 3. API-Driven Dashboard with Auth Build a dashboard that pulls real data from a public API, has user authentication, role-based views, and handles loading states, error states, and empty states gracefully. Not just the happy path. Every edge case. Stack: React or Next.js + Auth0 or Clerk + any public API + Tailwind. 4. Progressive Web App (PWA) That Works Offline Build a useful app that installs on mobile and works without internet. A note-taking app, a recipe saver, or a habit tracker. Handle offline caching, background sync, and push notifications. Stack: React or Vue + Service Workers + IndexedDB + Workbox. 5. AI-Powered Frontend Feature Build a feature that integrates AI into the UI. Smart search with semantic results, an AI writing assistant embedded in a text editor, or a chatbot that pulls context from the page. Show you can build the interface layer for AI, not just call an API. Stack: React or Next.js + OpenAI or Claude API + streaming responses + clean UX for loading states. 3 more ideas in the comments. Build. Break. Optimize. Prove. That's the story your portfolio needs to tell. Your tech stack list isn't proof. Your shipped, tested, accessible, performant code is. Your proof of work is your new resume. Share this with someone breaking into frontend engineering. Follow Yogi Gnanavel for career strategies to stay ahead.

46

23 Comments

US founders hiring Flutter in US time zones: your first paragraph decides who applies. If it's requirements and years, US seniors read it and think: "no chance." Lead with ownership instead: What ships in the first 90 days. What they own in production from day one. What they’ll have the authority to say no to. That’s the difference between a team add and a first mobile hire.

15

2 Comments

I had lunch with a developer (with $100,000,000+ AUM), and asked him what separates successful first-time developers from those who stumble.  His answer? Execution. Here’s what he emphasized: 1. Run daily stand-up meetings: - Every morning the project is active.  - Identify the day’s roadblocks.  - Keep the critical path visible. - Surface issues early. This is how small problems get solved before they turn into schedule risk. 2. Maintain a rolling punch list. - Walk areas as they finish. - Document issues immediately. - Close items while trades are still mobilized. Quality control alone can prevent weeks of delays. 3. Curate team accountability. - Punch walks aren’t just for the superintendent. - Everyone walks the site. - Everyone sees the issues. - No one hides behind scopes or emails. Quality improves quickly when responsibility is shared in the open. P.S. What would you add to the list?

86

43 Comments

I’ve been a SWE for the last 7+ years of my life. If I had to start all over again, I wouldn’t really invest in a course or resources without exploring by myself first. Here are 33 free roadmaps you can use today to start in any field in software engineering & crush it! 1. Backend Roadmap: https://roadmap.sh/backend & Frontend Roadmap: https://lnkd.in/dfk4-pa9 2. Full Stack Roadmap: https://lnkd.in/dRjCeDde 3. JavaScript Roadmap: https://lnkd.in/dA4ucqzz 4. TypeScript Roadmap: https://lnkd.in/dYdd4te4 5. React Roadmap: https://roadmap.sh/react 6. React Native Roadmap: https://lnkd.in/dBG72tqu 7. Vue Roadmap: https://roadmap.sh/vue 8. Angular Roadmap: https://roadmap.sh/angular 9. Node.js Roadmap: https://roadmap.sh/nodejs 10. CS Fundamentals Roadmap: https://lnkd.in/d7iDtkm6 11. GraphQL Roadmap: https://roadmap.sh/graphql 12. Rust Roadmap: https://roadmap.sh/rust 13. Go Roadmap: https://roadmap.sh/golang 14. C++ Roadmap: https://roadmap.sh/cpp 15. Java Roadmap: https://roadmap.sh/java 16. Spring Boot Roadmap: https://lnkd.in/dYRPibvp 17. ASP.NET Core Roadmap: https://lnkd.in/dfFXSAgZ 18. Software Architect Roadmap: https://lnkd.in/dDtNN-ch 19. Software Design and Architecture Roadmap: https://lnkd.in/dUHYkH95 20. System Design Roadmap: https://lnkd.in/d3S79YVJ 21. PostgreSQL Roadmap: https://lnkd.in/d8aTRkBD 22. SQL Roadmap: https://roadmap.sh/sql 23. MongoDB Roadmap: https://roadmap.sh/mongodb 24. AI and Data Scientist Roadmap: https://lnkd.in/dv7Tj4ft 25. Python Roadmap: https://roadmap.sh/python 26. Prompt Engineering Roadmap: https://lnkd.in/diZ2U4dG 27. DevOps Roadmap: https://roadmap.sh/devops 28. Docker Roadmap: https://roadmap.sh/docker 29. Kubernetes Roadmap: https://lnkd.in/dVGxcMKQ 30. Android Roadmap: https://roadmap.sh/android 31. Flutter Roadmap: https://roadmap.sh/flutter 32. UX Design Roadmap: https://lnkd.in/dWXzZgqu 33. Design System Roadmap: https://lnkd.in/dasuSJfZ

857

86 Comments

Looking for a software job? You NEED to read this. 8 things I would do if I were looking for a SWE job right now: Let's not sugarcoat it. The job market is absolutely brutal right now. I wish I could work with everyone 1:1, but I can't. So, this is for everyone I'm not able to coach: 1️⃣ 𝗦𝗽𝗲𝗰𝗶𝗮𝗹𝗶𝘀𝘁𝘀 𝗴𝗲𝘁 𝗵𝗶𝗿𝗲𝗱 𝘄𝗮𝘆 𝗳𝗮𝘀𝘁𝗲𝗿 𝘁𝗵𝗮𝗻 𝗴𝗲𝗻𝗲𝗿𝗮𝗹𝗶𝘀𝘁𝘀. This is the first thing I work on with every one of my clients. Get clear on what you're best at, and focus your materials around it. If you're not clear on: 1. What you do 2. Why it's valuable 3. Who you do it for Employers will pass over you. 2️⃣ 𝗢𝗽𝘁𝗶𝗺𝗶𝘇𝗲 𝘆𝗼𝘂𝗿 𝗟𝗶𝗻𝗸𝗲𝗱𝗜𝗻 𝗣𝗿𝗼𝗳𝗶𝗹𝗲 Your LinkedIn profile is more than an online resume. It's your personal landing page. Expect every recruiter and hiring manager to check your LinkedIn an interview. If it looks sloppy, they might reconsider their decision. (Free guide in the comments) 3️⃣ 𝗥𝗲𝘀𝘂𝗺𝗲𝘀 𝘀𝗵𝗼𝘂𝗹𝗱 𝗯𝗲 𝘀𝗸𝗶𝗺𝗺𝗮𝗯𝗹𝗲 Recruiters are skimming resumes for 7 seconds. It can't be full of fluff. Every word needs to earn its place on your resume. Add keywords where it makes sense, but don't keyword stuff. Focus on outcomes more than day to day responsibilities. Most of resume optimization comes from trimming fat, not adding content. 4️⃣ 𝗧𝗿𝗮𝗰𝗸 𝘆𝗼𝘂𝗿 𝗶𝗻𝗽𝘂𝘁𝘀 𝗮𝗻𝗱 𝗼𝘂𝘁𝗽𝘂𝘁𝘀 Most job seekers have no concrete idea how much effort they put in. Many have a rough idea, but it's a feeling more than solid data. Track your actions to review at the end of each week. Without data and reflection, you can not improve. 5️⃣ 𝗘𝗻𝗴𝗮𝗴𝗲 𝗼𝗻 𝗟𝗶𝗻𝗸𝗲𝗱𝗜𝗻 LinkedIn isn't just a job board. Invest time in actually using the platform. Post. Reshare. Comment. DM. Don't wait for opportunities to come to you. Force opportunities with your activity. 6️⃣ 𝗕𝘂𝗶𝗹𝗱 𝗿𝗲𝗹𝗮𝘁𝗶𝗼𝗻𝘀𝗵𝗶𝗽𝘀 𝗯𝗲𝗳𝗼𝗿𝗲 𝘆𝗼𝘂 𝗻𝗲𝗲𝗱 𝘁𝗵𝗲𝗺 The biggest networking mistake job seekers make is asking strangers for favors. Before you ask someone for a favor, turn them into an advocate. Build your network like you're building new friendships. If you can, do them a favor before asking for one back. 7️⃣ 𝗣𝗿𝗼𝘁𝗲𝗰𝘁 𝘆𝗼𝘂𝗿 𝘁𝗶𝗺𝗲. Your time is your most valuable asset in the job search. Most people waste their time because they don't know what they should focus on. Don't be like most people. Invest time into daily planning and reflections. It's worth the effort. Trust me. 8️⃣ 𝗗𝗼𝗻'𝘁 𝘁𝗿𝘆 𝘁𝗼 𝗱𝗼 𝗶𝘁 𝗮𝗹𝗹 𝗮𝗹𝗼𝗻𝗲 Surround yourself with a support system to stay sane. Friends. Family. Other Job Seekers. A mentor or coach. There's no reason to make the job search harder by doing it alone. Start taking back control of your search in this crowded market. You got this. I'm rooting for you. P.S. Any tips you want to add?

45

26 Comments

React.memo is quietly hurting a lot of React apps. Yes. Hurting. I’ve seen senior engineers do this: export default React.memo(Component) And then: <Component data={{ dark: true }} /> That object is recreated. On every render. React does a shallow comparison. New reference = re-render. So your “optimization” just became decoration. Now multiply that across 40–50 components in a dashboard. That’s hundreds of unnecessary renders per interaction. And here’s the twist: Overusing useMemo and useCallback can also slow things down. Every memo: Allocates memory Tracks dependencies Adds cognitive load Increases debugging complexity If a component renders in 0.2ms, let it render. Optimization without profiling is guessing. The real rule: → Expensive subtree? Memoize it → Passing unstable objects/functions? Stabilize them → Small presentational component? Don’t touch it Stop cargo-culting React.memo. Start understanding reference equality. What React “best practice” did you unlearn as you got more experienced? #React #Frontend #JavaScript #WebDev #Performance

24

3 Comments

Getting hired in crypto is not about luck. It is about proof. Recruiters are drowning in applicants right now. Many do not have real experience. Some are not even real people. If you want to land a role in this market, here is how to stand out: • Be verifiable → keep your LinkedIn active, show a track record (GitHub, Substack, onchain history, even Twitter threads). • Show you live it → recruiters can tell in 30 seconds if you have ever touched a wallet, bridged an asset, or traded onchain. • Communicate clearly → crypto teams are remote, async, global. If you cannot write or speak clearly, you are behind. • Contribute first → open source PRs, Discord answers, governance proposals. You do not need permission to start proving value. The bar is not “sound legit in an interview.” The bar is “prove you can add value on day one.” Crypto is competitive, but it is also open. There are no gatekeepers stopping you from showing you belong here.

239

51 Comments

I told a founder to STOP hiring senior engineers. He was about to spend his entire $500K pre-seed round on a team of 8 senior developers. Six months later, he's at product-market fit with 50% of his funding still in the bank. Let me explain why this was the right call: Pre-PMF startups face a brutal reality: 9 out of 10 products fail because they run out of money before finding what customers want. The founder planned to use all senior engineers to "build it right the first time." But here's the counterintuitive truth most technical leaders know but rarely say out loud: Before PMF, CODE QUALITY IS THE WRONG METRIC. Speed of learning is all that matters. The best senior engineers understand this deeply. In early-stage startups, your primary skill should not coding; it's knowing when NOT to use my full technical toolkit. Instead of an all-senior team, we recommended: - 1 experienced senior tech lead who understood startup constraints - 2 mid-level developers comfortable with rapid iteration - Clear guardrails on what to optimize for NOW vs. LATER By their Series A, they'll absolutely need more senior engineers to fix for scale. The tech debt is PERFECTLY FINE. Because the alternative was a beautifully architected product that nobody wanted. As Jeff Bezos says: Most decisions should be made with somewhere around 70% of the information you wish you had. If you wait for 90%, you're being slow. Senior engineers are invaluable; at the right stage. The best ones know exactly when their full powers should be deployed and when they should be deliberately constrained. #AAcareerMGT #recruiting #tech

55

13 Comments

I built a full-stack framework. Here's why. Every time I started a new Go + React project, I'd spend days on the same boilerplate: → Setting up auth → Wiring API routes → Creating admin panels → Writing the same CRUD handlers → Syncing types between Go and TypeScript So I built Grit. Grit is a batteries-included framework that fuses Go (Gin + GORM) with Next.js in a monorepo. One command scaffolds everything — API, frontend, admin panel, auth, Docker, database browser. But the magic is in code generation: $ grit generate resource Product --fields "name:string,price:float,published:bool" That single command creates: ✅ Go model with GORM tags ✅ CRUD handler with pagination, search, sorting ✅ React Query hooks ✅ Zod validation schemas ✅ TypeScript types ✅ Admin panel page with data table + form ✅ All wired together automatically Change a Go struct → TypeScript types and Zod schemas stay in sync. Automatically. It ships with auth, file storage (S3/MinIO), email, background jobs, cron, Redis caching, AI integration, and a visual database browser. The whole stack: Go + Gin + GORM → Next.js 16 → Tailwind + shadcn/ui → PostgreSQL → Redis → Docker It's open source. 📖 Docs: https://lnkd.in/e9gme798 💻 GitHub: https://lnkd.in/eQBgh6_W 📖 Cheatsheet: https://lnkd.in/dfKA6R5P 🎬 Demo Video: https://lnkd.in/d5QCbMZd If you're tired of wiring the same boilerplate every project, give it a try. I'd love your feedback. #golang #react #nextjs #fullstack #webdevelopment #opensource #framework #grit #typescript #developer

35

1 Comment

As agents get better at vibe coding, it's getting clear to me that they can generate tons of code, but they still need solid primitives to plug into. When I'm building something, I reach for Stripe for payments, Clerk for auth, Supabase for my database. Clean APIs, clear docs, they just work. Turns out agents need those same primitives. If you're having an agent generate a SaaS app, you point it at the same stack—APIs with clear boundaries, SDKs that make sense, connections that don't break. What's missing: REST APIs don't expose constraints or permission boundaries. An agent can read the docs, but can't ask "can I refund this charge given the user's current state?" without just trying it and seeing what breaks. I think as this pattern matures, these services will start optimizing specifically for agent consumption. Not just APIs, but schemas, capability metadata, example flows, etc. Everything to make integration safer and more predictable. Some are already shipping with MCP servers by default, turning every core primitive into something an agent can actually reason about and use securely. But here's the question: are we just teaching agents to replicate our architectural choices? Stripe takes 3% of every transaction. Clerk locks you into their auth patterns. These are human trade-offs based on human constraints. Maybe optimizing these services for agents is the right move. Or maybe the real opportunity is agents that know when NOT to use Stripe—when a simpler payment flow fits, or when rolling custom auth actually makes more sense. We might be automating our own assumptions.

27

8 Comments