Writing
Building in public.
Notes on AI integration, full-stack engineering, and what I learn shipping production code for clients.
Why Your LLM App Needs a Graph
Stateless chatbots are a dead end. Here is how I am using graph-based architectures to build reliable AI agents that can actually reason.
Why startups hire a senior developer in Bangkok
Timezone overlap with Europe, US mornings, and APAC. Senior rates without SF/London pricing. And someone who ships while you sleep.
Why I'm Ditching Traditional Serverless for the Edge
Cold starts and opaque pricing are killing developer joy. Here is why I am moving new workloads to Cloudflare Workers and keeping the heavy lifting in containers.
Why I'm Betting on AI Agents Over Chatbots
The chatbot interface is a dead end for complex workflows. Here is why autonomous agents using tools like LangChain and Vercel SDKs are the future of backend architecture.
The AI Agent Stacking Problem: Why Your LLM Calls Are Failing in Production
Moving from a working prototype to a reliable AI agent requires solving the complex issues of state management, structured outputs, and deterministic workflows.
The AI Agent Stack Isn't Ready, But We're Building It Anyway
Moving AI chatbots from demos to production requires a robust architecture. Here is the stack I am betting on for stateful agents, memory, and reliable tool execution.
The AI Agent Stack is Broken (And How We're Fixing It)
Most developers are over-engineering their AI agent infrastructure. Here is why I'm moving away from complex chains and back to simple, stateless functions.
The 'AI Agent' Stack is a Lie (and Why You Still Need It)
Most companies are building AI agents wrong, treating them like glorified chatbots wrapped in expensive infrastructure. Here is how I build agents that actually ship code.
Technical Co-founder for Hire: What to Look For
Need a technical co-founder for hire? Learn what to look for, from equity splits to tech stack choices (Next.js, Supabase). Avoid the 'rent-a-CTO' trap.
Supabase Auth + Expo: The Gotchas Nobody Tells You
Struggling with Supabase Auth in Expo? Learn the 5 critical gotchas involving session management, deep linking, and native modules that cause production crashes.
Supabase as Your Entire Backend in 2026
Most Node.js backends do five things — auth, data, real-time, serverless functions, and file storage. Supabase handles all five.
Why we use Supabase as our entire backend
No Express. No NestJS. Just Supabase — auth, database, real-time, edge functions, and RLS.
Stop Wasting Money on Cloud Cost Optimization
Before you hire a consultant to tune your AWS Reserved Instances or install another cost-monitoring agent, look at your actual architecture. The most expensive bill you have is the one generated by over-engineering.
Stop Building AI Chatbots, Start Building Agents
Most developers are still building glorified chatbots. Here is how I moved from simple prompts to autonomous agents using tool-calling and OpenAI's Assistants API.
Stop Building AI Agents, Start Building Features
AI agents are the shiny new toy, but in production, they are a nightmare. Here is why I prefer deterministic workflows over autonomous agents for 99% of client projects.
Stop Building Agents, Start Building Tools
Why the future of AI isn't autonomous robots, but deterministic tool-calling that integrates directly into your existing full-stack architecture.
Why Your Startup Needs an AI Audit Before Building
Stop burning your budget on unnecessary AI infrastructure. Learn why a technical AI audit saves 6 months of development and prevents technical debt before you write a single line of code.
The solo agency model: why I cap at three clients
No junior devs, no account managers, no hand-offs. Here's why working with one senior engineer beats a team of five.
Fixing silent iOS auth failure with PKCE
The auth flow that works everywhere except iOS WebViews. How PKCE fixes it.
How I Ship Full-Stack Apps Solo with AI Tools
Discover how I build scalable SaaS products in days, not months. Using Next.js, Supabase, and AI to replace a 5-person dev team. Read my full stack guide.
Self-hosting n8n on a VPS: My Automation Stack
Why I self-host n8n workflow automation instead of paying for the cloud tier — and the honest resource cost nobody talks about.
React Native vs Flutter in 2026: A Real Comparison
A senior engineer's breakdown of React Native vs Flutter in 2026. Real performance metrics, code examples, and architectural insights to help you choose the right stack.
Building a RAG Pipeline That Works in Production
Struggling with RAG latency and hallucinations? Here is how I architect production-grade Retrieval Augmented Generation systems using Next.js, Supabase, and OpenAI.
Monorepo architecture: 4 apps, 1 Supabase instance
How we structure a monorepo with iOS, Android, web, and admin sharing one Supabase backend.
On-demand LLM translation inside a chat interface
Real-time Thai ↔ English translation using LLMs with sub-second latency.
How I integrate AI into existing products without rewriting anything
Most AI features don't need a ground-up rebuild. Here's how I bolt intelligence onto products that are already in production.
When to Hire a Fractional CTO vs Full-Time: The Cost Breakdown
Deciding between a Fractional CTO and a full-time hire? I break down the costs, tech stack trade-offs, and timing for startups. Read the full guide.
Deploying Next.js on Cloudflare Pages vs Vercel in 2026
After running production workloads on both platforms for two years, here's the unvarnished comparison that actually matters for senior developers.
Building AI Agents That Actually Work (And Don't Hallucinate)
Moving beyond chat demos to reliable AI agents requires strict tool governance, structured outputs, and a 'human-in-the-loop' architecture.
Bangkok vs Eastern Europe for Your Dev Team: Cost & Speed
Comparing Bangkok vs Eastern Europe for your dev team? I break down the real costs, speed, and tech stack quality (Next.js, Supabase) to help you decide.
Why I Moved from AWS to Cloudflare (and When You Should Too)
I spent 12 years building on AWS. Last year, I migrated my core stack to Cloudflare. Here is the deep dive on latency, costs, and the architecture shift.
My AI-augmented dev workflow in 2025
Claude Code, Cursor, n8n agents, and local LLMs. Here's exactly how I use AI tools to ship 3x faster.
How Much Does AI Integration Cost in 2026? Real Breakdown
Planning your 2026 AI roadmap? A senior engineer breaks down real costs for OpenAI API, RAG, and custom agents. Stop guessing and start budgeting.
AI agents in production: automating AML screening for a FinTech
How we built AI agents that screen against 1.7M+ sanctions records in seconds. Not a demo — it's processing real compliance cases.