About
I have already experienced working on both small to enterprise-grade software products. I…
Activity
1K followers
-
Rhaidzsal Ali reposted thisRhaidzsal Ali reposted thisThe more you learn about AI, the more there is to do. Most people think this is AI failing to deliver its promise. That's far from what's happening. A lot of people i've met are in full catch-up mode with AI... mostly excited, some worried, maybe exhausted, or all at the same time. Wrote about mindset that can help.You're not on a Learning Curve. You're on a Treadmill.You're not on a Learning Curve. You're on a Treadmill.
-
Rhaidzsal Ali reposted thisRhaidzsal Ali reposted thisI stopped testing APIs one endpoint at a time. For years my workflow looked the same regardless of the tool. Curl, Postman, whatever. Hit an endpoint, check the response, copy the ID, paste it into the next request, repeat. Works fine for checking if a single endpoint returns 200, but tells you nothing about whether your API works as a flow. A user doesn't call POST /users and stop. They sign up, update their profile, change their email, maybe delete the account. That's five endpoints depending on each other, and I was spending more time chaining them manually than thinking about what could break. Now I describe the whole scenario in plain text: "Register a user. Update their email. Fetch the profile, confirm it changed. Try the same email again, should fail. Delete. Fetch again, should 404." An AI agent runs the chain, passes data between steps, and reports where it broke. Since I switched, I've been catching things I never found before. Race conditions between create and update. Broken state after partial failures. Pagination behaving differently after a delete. These bugs live between endpoints, and you only see them when you test the full sequence. A friend in QA picked this up too. He used to wait for someone to explain the API before he could test anything. Now he writes scenarios in plain language and gets results in seconds. Another friend on frontend stopped hardcoding mock responses and just checks real API behavior before building components. Fewer integration bugs, zero extra setup. Endpoints are implementation details. Users think in flows. Testing should too. What's your go-to way to test API chains end to end? #BackendDevelopment #APITesting #DeveloperProductivity #GoLang #DevTools
-
Rhaidzsal Ali reposted thisRhaidzsal Ali reposted thisWill AI completely replace software engineers soon? My answer is purely mathematical, and it will change how you think forever. A single engineer makes roughly 10 micro-decisions every single working day. React or Vue? Monolith or microservices? Encapsulation or inheritance? HTTP or gRPC? That's 220 micro-decisions made by one engineer across a full working month. AI is incredibly accurate, let's generously call it 99% accurate right now. But 1% still means 2 wrong decisions slipping through every single month. Now here's where it gets interesting, and slightly terrifying, to think about. If those 2 wrong decisions happen on day one, your product is completely broken forever. The entire rest of the build compounds on a fundamentally flawed, unstable foundation. If they happen at the end, someone still has to find and fix them. And that someone needs serious engineering judgment to prevent cascading downstream errors. We actually tested this theory with a real company trying to build fully AI. We tried to build an entire SaaS platform using only AI-generated code. The code worked initially, but nobody could actually maintain it afterwards. You couldn't build on it. You had to scrap everything and fully rewrite it. AI can write code. But it cannot yet replace engineering judgment at scale. The 1% doesn't sound scary until it's sitting at the foundation of everything. #AI #softwaredevelopment
-
Rhaidzsal Ali reposted thisRhaidzsal Ali reposted thisOpenClaw Cebu happened tonight! 🦞 The OpenClaw PH community is moving fast and it's been fun watching it come together city by city. Shoutout to Zero-Ten Park Philippines (formerly The Company Philippines) for hosting and Symph for snacks. We had fellow claw builders share what they've been working on: David Marquez from OpenClaw Manila, Raven Duran from OpenClaw CDO, Jerel Velarde and Prince John Santillan from the Cebu tech community! OpenClaw CDO is next week: https://luma.com/ej1jr2ip
-
Rhaidzsal Ali reposted thisRhaidzsal Ali reposted thisServer-Sent Events use 60% less memory than WebSockets. And OpenAI chose them over WebSockets for ChatGPT streaming. 95% of "real-time" features only need server → client updates. Dashboards. Notifications. Live feeds. AI responses. Log streams. Yet every team defaults to WebSockets — and pays the tax: // WebSocket: protocol upgrade + manual reconnection const ws = new WebSocket('ws://example.com/socket'); ws.onclose = () => { // YOU write exponential backoff // YOU handle reconnection attempts // YOU manage heartbeat pings }; // SSE: 3 lines, auto-reconnect built in const es = new EventSource('/api/stream'); es.onmessage = (e) => updateUI(JSON.parse(e.data)); // Browser handles reconnection automatically Here's what the benchmarks show (Timeplus, 2024): • Max throughput: 3M events/sec — identical for both • Latency difference: 3ms (lost in network noise) • Memory at 10K connections: SSE ~20MB vs WebSocket ~50MB The real killer: SSE is just HTTP. No proxy issues. No firewall blocks. No special CDN config. curl works for debugging. When this doesn't apply: • Multiplayer games (constant bidirectional traffic) • Collaborative editing (Google Docs-style conflict resolution) • Video call signaling (WebRTC ICE negotiation) Shopify streamed 323 billion events during Black Friday 2022 using SSE — zero WebSocket complexity, sub-300ms latency globally. What real-time feature are you building? Chances are SSE is all you need. #ServerSentEvents #RealTime #EventSource #Streaming
-
Rhaidzsal Ali shared thisNow this is great! 👀Rhaidzsal Ali shared thisCloudflare just shipped Vinext, a clean-room reimplementation of the Next.js App Router API surface using Vite as the build engine. The headline numbers: 4.4x faster production builds and 57% smaller client bundles (168.9 KB down to minuscule 72.9 KB gzipped). Where the reigning OpenNext reverse-engineers next build output (which breaks whenever Next.js changes its internals), Vinext reimplements the API contract itself. It's a proper replacement for those looking to completely break away from Next.js (and more importantly, Vercel). The whole thing was built by a SINGLE engineer directing Claude through 800+ sessions in about a week, $1,100 in API tokens. Reading for humans (but mostly devs): https://lnkd.in/ef_a6fxX
-
Rhaidzsal Ali reposted thisRhaidzsal Ali reposted thisCloudflare just shipped Vinext, a clean-room reimplementation of the Next.js App Router API surface using Vite as the build engine. The headline numbers: 4.4x faster production builds and 57% smaller client bundles (168.9 KB down to minuscule 72.9 KB gzipped). Where the reigning OpenNext reverse-engineers next build output (which breaks whenever Next.js changes its internals), Vinext reimplements the API contract itself. It's a proper replacement for those looking to completely break away from Next.js (and more importantly, Vercel). The whole thing was built by a SINGLE engineer directing Claude through 800+ sessions in about a week, $1,100 in API tokens. Reading for humans (but mostly devs): https://lnkd.in/ef_a6fxX
-
Rhaidzsal Ali reposted thisRhaidzsal Ali reposted thisDon’t use clever operators. Normal IFs are classics that win. Developers tend to employ fancy one-liners. This puts the implementation of different software behavior in one line. It’s fine, until prettier automatically formats it into a multi-liner 🤯 ``` return !!getSenderFromPayment(payment) ?? isSenderOwnAccount(payment, accounts) ?? !isLikeDefaultSender(payment); ``` This is hard to read. And easy to miss the test for. Well, there are 4 logical branches hidden behind this code. Each represented a different software behavior. The fix? Normal IF conditions 🙌 It’s nothing fancy, nor clever. But does the job. Let’s leave the optimization job to a compiler. The AI-generated code is not an exception. Clarity > fancy/clever Repost ♻️ (to help your network learn, too) Do you document this kind of design pattern in your main AI instructions file or in a skill?
-
Rhaidzsal Ali reposted thisRhaidzsal Ali reposted thisAI coding has ridiculously improved the past few months... I'd say since January, my code has been 99% AI generated and it updates everything flawlessly. The planning mode is perfect for larger, complex change. I tend to over prompt, and would prompt even simple deletion of print debug lines... And even when it takes longer for AI, than me just deleting/changing the code. Until I started refactoring our code 3 days ago purely with prompt in planning mode, the result is catastrophic. Poor, unreadable code, and overengineered design. Original 400 lines become 700 lines. And making it super hard to read... It's like AI writing code for AI. So, today I decided to revert back and manually refactor our codebase. Reviewing line by line, result is a work of art. Lesson learned? AI is extremely good at generating code that satisfies a request. It is less good at generating code that is pleasant to live with. Refactoring is a taste problem. It requires judgment like: “This is simpler even if it’s less ‘clean’ by textbook standards.” or “We don’t need this pattern here.” or "I don't want backward compatibility, I want clean code." Those decisions are deeply contextual and they rely on a mental model of your team, and human tolerance for complexity.
Rhaidzsal Ali commented on a post
3d
killua and nanika hahaha
-
Rhaidzsal Ali liked thisRhaidzsal Ali liked thisWow, I just learnt a new way of writing tests in JavaScript Just discovered that Vitest has support for "in-source testing" — you can write your tests alongside source code in the same file, strip them out during build. Was reading some Rust code and thought that Rust's `#[cfg(test)]` was really neat as it lets you have tests alongside source code in the same file, no redundant exporting of implementation details just for testing. Then I found out Vitest had this feature since years ago. Yes I'm slow, but better late than never. That's why it pays to read the docs every now and then, you never know what you can find. ——— ♻ Repost to help others discover 📕 Save the post so you don't miss it 💡 Follow me Yangshun Tay and my company GreatFrontEnd for more
-
Rhaidzsal Ali liked thisRhaidzsal Ali liked thisStop creating external variables for resolve and reject. 🔌 We’ve all been there: declaring "let resolve" outside, creating a Promise, and assigning the function inside the callback. In 2026, that’s just unnecessary noise. THE EVOLUTION OF DEFERRED PROMISES: 📍 The "Manual" way (Boilerplate): let resolve, reject; const promise = new Promise((res, rej) => { resolve = res; reject = rej; }); 📍 The 2026 way (Native & Clean): const { promise, resolve, reject } = Promise.withResolvers(); WHY IT’S AWESOME: 🔹 Less Boilerplate: No more extra wrappers or mixing up execution contexts. You get everything you need in one line. 🔹 Readability: Your code becomes linear and understandable at a glance. No more "searching" for where the promise is actually resolved. 🔹 Pattern-friendly: It’s the perfect tool for wrapping legacy event-based APIs or streams into modern async/await patterns. It’s a tiny feature that makes code significantly cleaner. If the platform gives you a specialized tool to handle deferred execution, it's time to retire the manual hacks. Have you made the switch to native .withResolvers() yet, or are you still "manually" resolving your promises out of habit? Let’s discuss below! 👇 #JavaScript #AsyncJS #CleanCode #Programming #WebDevelopment #SoftwareEngineering #CodingTips #TechTrends
Experience & Education
-
Titan FX
View Rhaidzsal’s full experience
See their title, tenure and more.
or
Licenses & Certifications
-
Electronic Data Processing Specialist in Computer Programming
Department of Information and Communications Technology - Philippines
Issued -
-
Volunteer Experience
-
Dev Guild Contributor
BanklessDAO
- 8 months
Science and Technology
Actively participating on the dev guild, specifically for the GovBot MVP.
-
OSS Contributor
KodaDot
- 2 months
Science and Technology
Contributing pull requests for features and issues to the open-sourced NFT marketplace on Kusama.
Courses
-
Bachelor of Science in Computer Science, Major in Software Technology (ST)
-
Honors & Awards
-
Cum Laude
Western Mindanao State University
Test Scores
-
Career Service Professional Eligibility Exam
Score: 81.28%
Passed an examination for Civil Service Eligibility.
Languages
-
English
Full professional proficiency
-
Filipino
Native or bilingual proficiency
-
Spanish
Elementary proficiency
-
German
Elementary proficiency
Recommendations received
4 people have recommended Rhaidzsal
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content