Eric Yi

We all know the mantra: "𝘕𝘰𝘥𝘦.𝘫𝘴 𝘪𝘴 𝘴𝘪𝘯𝘨𝘭𝘦-𝘵𝘩𝘳𝘦𝘢𝘥𝘦𝘥." But that’s only half the story. While the Event Loop runs on a single thread, 𝗹𝗶𝗯𝘂𝘃 (the C library powering Node’s async magic) keeps a secret weapon in the back: The 𝗪𝗼𝗿𝗸𝗲𝗿 𝗣𝗼𝗼𝗹. 🛠️ By default, Node.js assigns 𝟰 𝘁𝗵𝗿𝗲𝗮𝗱𝘀 to this pool. This is where the heavy lifting happens—file system operations, DNS lookups, and crypto functions like 𝚙𝚋𝚔𝚍𝚏𝟸. But what if 4 threads aren't enough? Or what if they are too many? Enter the magic world of 𝚄𝚅_𝚃��𝚁𝙴𝙰𝙳𝙿𝙾𝙾𝙻_𝚂𝙸𝚉𝙴 param. 🔬 𝗧𝗵𝗲 𝗘𝘅𝗽𝗲𝗿𝗶𝗺𝗲𝗻𝘁 I ran a benchmark using 𝚌𝚛𝚢𝚙𝚝𝚘.𝚙𝚋𝚔𝚍𝚏𝟸 to hash passwords. I threw 100 parallel tasks at Node.js to see how it handled the pressure. 𝗧𝗵𝗲 𝗖𝗼𝗱𝗲: (See the first comment for the snippet! 👇) 𝗧𝗵𝗲 𝗥𝗲𝘀𝘂𝗹𝘁𝘀: 🐢 𝗨𝗩_𝗧𝗛𝗥𝗘𝗔𝗗𝗣𝗢𝗢𝗟_𝗦𝗜𝗭𝗘=𝟭 Execution: 6.27s Why: The bottleneck is real. Tasks form a massive queue. 🚗 𝗗𝗲𝗳𝗮𝘂𝗹𝘁 (𝗦𝗶𝘇𝗲=𝟰) Execution: 1.75s Why: Better, but with 100 tasks, 4 workers are still swamped. 🏎️ 𝗨𝗩_𝗧𝗛𝗥𝗘𝗔𝗗𝗣𝗢𝗢𝗟_𝗦𝗜𝗭𝗘=𝟳𝟬 Execution: 1.26s 🚀 Why: We maximized parallelism (given the machine's capabilities). 💡 𝗪𝗵𝗲𝗻 𝘁𝗼 𝗧𝘂𝗻𝗲 𝗧𝗵𝗶𝘀? ✅ 𝗗𝗢 𝗜𝗻𝗰𝗿𝗲𝗮𝘀𝗲 𝗜𝘁 𝗜𝗳:  • You have many concurrent CPU-bound ops (Crypto, hashing, compression).  • Your app is I/O heavy (FS operations) and the default pool is starving.  • You have a machine with high core counts to support the concurrency. ❌ 𝗔𝗩𝗢𝗜𝗗 𝗜𝗻𝗰𝗿𝗲𝗮𝘀𝗶𝗻𝗴 𝗜𝘁 𝗜𝗳:  • The bottleneck is the network or external DB (JS can’t fix latency).  • You exceed your logical CPU cores significantly. Too many threads = Context Switching 📉. The CPU wastes time jumping between tasks rather than working on them. 🧠 𝗧𝗵𝗲 𝗧𝗮𝗸𝗲𝗮𝘄𝗮𝘆 Node.js defaults are great for general cases, but "Default" doesn't mean "Optimized" for your specific workload. Don't just trust the settings. Measure, benchmark, and tune. What’s your experience with Node performance tuning? Have you ever had to touch the 𝚄𝚅_𝚃𝙷𝚁𝙴𝙰𝙳𝙿𝙾𝙾𝙻_𝚂𝙸𝚉𝙴 in production? Let’s discuss in the comments! 👇 #NodeJS #BackendDevelopment #JavaScript #PerformanceTuning #SoftwareEngineering

43

7 Comments

Most FAANG interview prep advice posted online feels like a movie trailer. Cinematic. Over the top, and totally disconnected from real life. "Wake up at 4am." "Do an ice plunge." "Grind 3500 Leetcode problems." "Master 50 system design patterns in 1 weekend". "Land 6 figure roles today with 0 experience by doing my exclusive course." So much nonsense... It looks impressive on a post. But it’s not how people really prepare, especially when you have a full-time job, a life, and a limited amount of energy after long day of work. Here’s one realistic way you can prepare for tech interviews in 2025 (without burning yourself out): ➤ Data Structures & Algorithms (DSA) isn't about doing 3500 questions. You just need to understand the building blocks: → Arrays → Linked Lists → Stacks & Queues → Hashmaps → Heaps → Trees → Graphs → Tries → Matrices And the key patterns: → Two Pointers → Sliding Window → Binary Search → Recursion → Basic Dynamic Programming When the patterns click, the questions stop feeling like magic tricks. Aim for 60–90 minutes/day. Consistency > intensity. ➤ System design isn't about memorising diagrams You just need to focus on these core fundamentals: → API Design: rate limiting, versioning, REST basics → Databases: SQL vs NoSQL, indexes, transactions, strong vs eventual consistency → Caching → Scalability: Horizontal vs Vertical scaling, Load balancing, CDNs → Queues & async processing → Sharding & replication → Failure handling & retries → Observability (logs, metrics, tracing) This isn't an exhaustive list, but these pillars should carry you through most rounds. Remember, your reasoning matters more than your drawing skills. ➤ Behavioural interviews aren't less important than technical interviews Prepare STAR stories about: → Ownership → Customer Obsession → Conflict → Ambiguity → Learning something quickly → Leading without authority Take a look at Amazon's leadership principles, and see how your previous experiences matches up with the LPs. Real prep is slow, steady, boring, and unbelievably effective. There are no shortcuts. It takes time. If you focus on fundamentals, stay consistent, and take care of yourself along the way...you'll be in a good position. Learn. Build. Rest. Repeat. You've got this. If you want genuinely helpful resources, do check out: → System Design One with Neo Kim → Coding Challenges with John Crickett → Taro (YC S22) for behavioural interview preparation with Alex Chiou and Rahul Pandey. → Regular tips on LinkedIn: Anton Martyniuk, Daniel Moka, Demitri Swan and saed. #softwareengineering

3,242

102 Comments

If you're a Senior Engineer looking to move up, the next role will likely be a Staff Engineer. And in this guide, Shruti shares tips from her own experience of getting promoted to Staff Engineer at PayPal and Slack. She talks about what Staff Engineers do (and how it's different from Seniors), why you might not be getting promoted, and how to take that next step. https://lnkd.in/g58dnFEG

103

3 Comments

🚨 Microsoft just announced a major shift in its flexible work policy: starting February 2026, employees near Puget Sound will be expected to work 3 days per week in the office, with global rollout to follow. This is a fascinating signal in the ongoing tug-of-war between flexibility and structure. On one hand: ✅ In-person collaboration does spark creativity and speed in ways remote often can’t. ✅ Shared space builds culture, trust, and those “hallway moments” that shape teams. On the other: ⚠️ Many employees have built their lives around remote flexibility. ⚠️ The value of autonomy and reduced commute stress is hard to ignore. ⚠️ Hybrid done poorly risks being the “worst of both worlds” - travel without true collaboration. My take? The real win isn’t “3 days” or “5 days” or “remote forever.” It’s clarity + intentionality. ➡️ If in-office time is structured around collaboration, innovation, and culture → it creates value. ➡️If it’s just presence for presence’s sake → it erodes trust. Microsoft’s move will ripple across the industry. The companies that succeed won’t just copy a policy - they’ll design hybrid work that respects people’s lives while amplifying team outcomes. What do you think? Is the 3-day baseline the sweet spot, or just another rigid rule in disguise? #FutureOfWork #HybridWork #Leadership

16

2 Comments

Stop treating iOS and Android as separate projects. This mindset is costing you weeks of development time. I made this mistake for several years. Despite using React Native, I was still thinking in platform silos - separate build processes, separate testing, separate deployment schedules. I'd fix a bug on iOS, then remember I needed to check if it affected Android too. The breakthrough came when I realised React Native isn't just shared code - it's one product with two interfaces. When you embrace true cross-platform thinking, everything changes: → Shared CI/CD pipelines → Unified testing strategies → Simultaneous releases → One bug fix = both platforms covered The mental shift that changed everything: Stop thinking "iOS app + Android app." Start thinking "mobile app with two interfaces." Build one pipeline, one dashboard, one release process. Your development velocity will double overnight. The teams that get this ship faster, debug quicker, and maintain better code quality. The teams that don't... well, they're still debugging iOS while their Android users wait. Follow me for more React Native insights from 10+ years of mobile development.

22

1 Comment

🔥Senior Frontend Developer Interview Prep, Real Questions You Should Be Able to Answer in 2025 If you’re preparing for senior-level roles, these are the kinds of advanced, scenario-based questions companies are actually asking today, not textbook theory, but real-world engineering depth. 🧠 JavaScript & TypeScript (Advanced) • Explain event loop behavior with nested promises + setTimeout + async/await • Implement debounce, throttle, memoize, custom curry • How do you detect & fix memory leaks in a JS-heavy app? • Explain how garbage collection works in V8 • Write a deep clone function without lodash • Design a plugin system using closures • What happens internally when you call new? • Implement your own Promise.all / Promise.race / Promise.any • Explain TypeScript type narrowing • Write complex utility types (Partial, Pick, Record) from scratch ⚡ Angular (Senior-Level) • How to avoid unnecessary change detection in large Angular apps? • Convert a large component to OnPush + immutability — explain steps • Design a custom global state management system (without NgRx) • Build a dynamic form engine from API schema • Create a custom RxJS operator to track API performance • Explain micro-frontends in Angular: architecture, routing & shared services • Fix "ExpressionChangedAfterItWasChecked" — root causes • Optimize lazy loading + preloading for 10+ feature modules • Handle role-based permissions with guards + interceptors • Debug a production issue where Angular becomes slow after navigation 📦 Performance Optimization • Reduce bundle size by 30–40% — exact steps • Explain TBT, LCP, CLS and how to improve each • Apply tree-shaking, deferred loading & code splitting in SPA • When should you detach change detection? • Debug a CPU spike caused by infinite UI loop 🛠️ HTML, CSS & UI Systems • Build a responsive grid without frameworks • Explain CSS containment and performance impact • Build accessible UI components (WCAG) • Implement modal with focus trapping • Fix layout shift in image-heavy UI 🧩 System Design for Frontend • Design a StackBlitz-like online editor (frontend system design) • Build a notification system with offline support • Create micro-frontend architecture for enterprise apps • Secure frontend routes + tokens • Build a real-time dashboard with streaming data + charts 🧪 Hands-On Coding Rounds • Reusable table component with sort, filter, pagination, virtual scroll • Build your own switchMap using pure JS • Write a diff algorithm for two JSON objects • Implement drag & drop without libraries • Build a mini SPA router • Implement mini Redux store (dispatch, subscribe, reducer) • Custom pipe to filter deeply nested data efficiently • Build search-as-you-type with caching + debounce 🔥 If you can confidently explain and code these, you’re truly operating at senior-level frontend engineering. #frontend #javascript #reactjs #angular #webdevelopment #codinginterviews

40

Your coding agent keeps asking for clarification. Or gives you generic code that doesn't fit your project. The problem? Your prompts are too vague. Here's how to be more specific: Cline works best when your ask is specific. Vague requests like "fix this bug" or "make it better" give Cline nothing to work with. Missing context: • What should be working • What's actually happening • Where the problem occurs Result: Ambiguous prompts make troubleshooting nearly impossible. Here are 3 techniques that turn vague prompts into precise instructions: 1. ZERO-SHOT PROMPTING Ask Cline to perform tasks without providing examples. Works because modern language models have been trained on vast amounts of code and documentation. Example: "Write me a React app for my e-commerce company" Use when: You need well-established patterns and common development tasks. 2. ONE-SHOT PROMPTING Provide a single example of how you want something done, then ask for similar work. Example: "Write me a helper function for addition: function add(a, b) { return a + b; } Now write me a helper function for division." Bonus: Cline automatically explores your codebase and uses existing code as implicit examples. Use when: You have specific style requirements or existing patterns. 3. CHAIN-OF-THOUGHT PROMPTING Explicitly outline the steps you want Cline to follow when tackling complex problems. Instead of: "Fix this bug" Try this: "This function throws a null-pointer exception. 1. Check the build log for compile warnings 2. Look for runtime errors in terminal output 3. Examine source code for null variables 4. Review documentation before implementing fix" Use when: Complex, multi-step processes where sequence matters. The key: Match the right technique to each situation. Understanding these approaches helps you anticipate how Cline will interpret and act on your requests. More on prompt fundamentals here: https://lnkd.in/gHt3YzNR What's your experience with AI coding tools? Which prompting challenges do you face most often? Check out AI Coding University for more guides like this: https://cline.bot/learn

26

1 Comment

GenZ has access to more career advice than any generation before them. Yet, they're the most confused about what to do next. Here's the uncomfortable truth no one's talking about: Last week, a final-year engineering student reached out to me. Smart student, 8.5 CGPA, active on LinkedIn and follows all the "top career influencers." He told me: "Sir, I've consumed so much content. But I still don't know what I should do." The problem has nothing to do with lack of guidance. The real problem? Too much noise disguised as guidance. → Every YouTuber shares "5 steps to success" → Every Instagram influencer sells a different dream → Every LinkedIn post promises a shortcut But here's what they don't tell you: Most of them have never worked in the industries they're advising you on. They've mastered content creation, not career building. Let me make this simple for you. These 5 filters will save you years of confusion: 1. Check their track record, not their follower count → Have they actually built a career in the field they're teaching? → Do they have real-world experience or just content creation skills? 2. Look for depth, not just entertainment → Genuine experts explain the "why" and "how," not just the "what" → They give you frameworks, not just motivational quotes. 3. See if they customize advice or give one-size-fits-all solutions → Real mentors understand that your path is unique → Cookie-cutter advice works for views, not for your life. 4. Observe how they handle questions → Do they engage genuinely or just redirect you to paid courses? → Are they patient with "basic" questions or dismissive? 5. Trust people who've helped others succeed → Ask for testimonials, case studies, or real transformations → Results speak louder than polished Instagram reels. You don't need 100 voices telling you different things. You need ONE trusted guide who understands your situation. Someone who's walked the path, made the mistakes, and knows what actually works. My advice to every GenZ professional feeling lost: Stop consuming content for the sake of consumption and start filtering for credibility. The right mentor won't give you all the answers but they'll teach you how to find your own answers. What's been your biggest struggle in finding genuine career guidance? Drop a comment. Let's talk about it. CREDIT:Priyank Ahuja

34

Want to clear any Software Engineering role in 2025? Use these resources.. 13 GitHub repositories that will make Software Engineering interviews super easy for you! 1. Tech Interview Handbook Yangshun Tay (109k stars): https://lnkd.in/gpyq4H5p 2. System Design Primer by Donne Martin (252k stars): https://lnkd.in/g9KU_kSP 3. System Design 101 by Alex Xu (56k stars): https://lnkd.in/g77Z7iGH 4. Top System Design Resources by Gaurav Sen (14k stars): https://lnkd.in/e9tEQnt6 5. Best Company and Personal Engineering Blogs by kilimchoi (29k stars): https://lnkd.in/g7yUkd7r 6. The best developer Roadmaps by Kamran Ahmed (274k stars): https://lnkd.in/guF3KCvP 7. Computer Science Papers for System Design by Arpit Adlakha: https://lnkd.in/gpGzcTaN 8. All Resources for Coding, LLD and HLD interviews by Arpit Adlakha (1.2 k stars): https://lnkd.in/gdDk7yjm 9. Coding Interview University (252k stars): https://lnkd.in/ggcPaMph 10. System Design Questions/Solutions Collection by Ashish Pratap Singh (10k stars): https://lnkd.in/gtt_GZeB 11. What Happens when you type google.com in browser (Most Indepth Answer): https://lnkd.in/gbWkADQ6 12. 30 seconds of Code (119k stars): https://lnkd.in/gZMx_zHa 13. System Design Chapter Wise Explained (28k stars): https://lnkd.in/giindjpe Want to ace tech interviews? ✅ Build real-world projects (share on GitHub) ✅ Get a strong tech mentor who can guide & challenge you ✅ Practice mock interviews early to build confidence & improve performance ✅ Build your online presence (GitHub, LinkedIn, blogs, etc.) Start early. Stay consistent. Results will follow. 🔁 Save this post and tag a tech buddy who is prepping for interviews 👉 Follow Lakshmi Marikumar & Everyone Who Codes (EWC) for tech growth & job search strategies. We have a simple mission: helping engineers get hired! Reach out or DM me for any mentorship program for success in tech interviews! #systemdesign #softwareengineering #techjobs #interviews #backend #softwareengineer #softwaredeveloper #MAANG #2ndtier #hiring #jobsearch #mentorship #mockinterview #projects #onlinepresence

141

11 Comments

🦺 🛟 Agent Safety > Model Safety LLMs didn't break your app. Agents can. As we wire LLMs to tools, browsers, payments & code the risk shifts from bad text to bad actions. Top failure modes (seen in the wild & in papers): 🚩 Prompt/Content Injection -> Tool Misuse. Malicious web pages or data instruct agents to steal secrets or make unintended API calls. This is now a primary risk for browsing and MCP - style tool agents. 🚩 Over permissioned tools & long-lived tokens. Agents get "god mode" scopes. A single injection becomes account takeover or irreversible ops. Security notes for MCP emphasise least privilege and short-lived auth. 🚩 Unsafe web autonomy. Benchmarks show web agents will attempt harmful tasks unless constrained (posting misinformation, illicit sales, etc.). You must measure misuse not just assume guardrails. 🚩 Supply-chain & retrieval poisoning. Agents trust plugins, third-party tools & indexed data that an attacker can taint. OWASP’s newest GenAI Top 10 calls this out explicitly. 🚩 Process safety gaps. NIST's GenAI Profile (AI 600-1) warns that organizations ship agents without role clarity, human-in-the-loop or incident playbooks. ✅ Model safety reduces bad text. Agent safety prevents bad state. Treat the agent like a junior SRE - with narrow permissions, audited actions and clear escalation paths. If you're building agents, what's the one control you won't ship without? Also drop a comment if you'd like me to share a follow-up post on how to tackle each of these risks. #AgenticAI #AISafety #AIAgents #GenAI #CyberSecurity #MCP #OWASP #NIST #AITrust #ResponsibleAI #AIForBusiness

5

🧠 JavaScript Contractors: How to Prep for Your Next Interview Let’s be honest—it’s a client-driven market right now. That means clients have the power to be selective, and contractors need to be sharp, fast, and ready to deliver value from day one. Whether you're stepping into a short-term gig or a long-term contract, interviews are your moment to show you're not just fast, but smart. Focus on the Fundamentals: 1. Data Structures & Algorithms: Brush up on arrays, maps, recursion, and sorting. 2. Mathematics: Useful for animations, logic, and edge-case handling. 3. Object-Oriented Programming: Think modular, reusable, and maintainable code. 4. API: Know how to design, consume, and debug REST/GraphQL endpoints. 5. Multi-threading Concepts: Understand async/await, event loop, and web workers. 6. Testing: Be ready to talk about unit tests, integration tests, and tools like Jest or Cypress. 🛠️ Practice in Context Use platforms like LeetCode, CodeWars, or even GitHub issues to sharpen your skills. But also rehearse how you’d apply these in real-world JS projects. especially in React, Node.js, or TypeScript environments. 💬 Communicate Clearly Contract roles move fast. Show you can explain your decisions, collaborate well, and hit the ground running. 🎯 Final Tip Prep like you're solving problems for a client, not just passing a test. That mindset makes all the difference. #JavaScript #Contractors #InterviewPrep #FrontendDev #NodeJS #TechCareers #CodingTips #SoftwareEngineering

15

4 Comments

🤯 The coding agents are so good now, I think it actually is feasible to use them as interview baseline for SWE hiring. Set up a project repo and 3 actual tasks, if a developer can't do better than coding agent (e.g. Claude Code CLI), don't hire them.

30

13 Comments