Product4 min read
Metrics That Matter Early Stage
Vanity metrics vs. real signal. What I track for AssessAI and why most startup dashboards are performance art.
Blog
48 postsLong-form articles and short thoughts on AI, product, and engineering.
AI & LLM5 min read
The Context Window Trap
1M token context windows don't mean you should use them. When to chunk, summarize, or restructure instead of stuffing everything in.
System Design9 min read
System Design: Distributed Job Scheduler
Designing a cron-at-scale system — priority queues, exactly-once execution, retry with dead letter queues, and the monitoring that keeps it honest.
Thoughts1 min read
Remote work is a skill
Working across Japan, USA, and India taught me that async communication is everything.
Software Dev4 min read
Git Workflow for Solo Founders
Trunk-based dev, feature branches, conventional commits. What works when you're the only person pushing code.
AI & LLM6 min read
LLM Evaluation Is Hard
Building LLM-as-a-Judge for AssessAI taught me why automated evaluation needs human calibration — and where it breaks down.
Product5 min read
Building AssessAI: Why Coding Tests Are Broken
Most companies still use LeetCode to hire senior engineers. In 2026. When AI does 70% of the coding. Here's what I built instead.
AI & LLM5 min read
Voice AI in Production
What it takes to deploy conversational Voice AI that handles real phone calls. Latency budgets, failure modes, and lessons from Avoca.
Tools3 min read
My Claude Code Setup: 68 Commands, 15 Agents, 86 Skills
How I turned Claude Code into a complete engineering team in a terminal.
Product4 min read
Solo Founder Stack 2026
Next.js + Supabase + Vercel + Claude Code. The stack that lets one person ship like five.
System Design9 min read
System Design: File Storage Service
Designing S3-like object storage — chunking, deduplication, CDN integration, and the metadata layer that ties it all together.
AI & LLM4 min read
Embedding Models Compared: Cost, Quality, Latency
OpenAI ada-002 vs text-embedding-3 vs Cohere vs local models. Real benchmarks from production retrieval systems.
Thoughts1 min read
Model size is a vanity metric
Ran Kokoro-82M on my MacBook for TTS. 337MB. Broadcast-quality speech. Zero cost.
Tools5 min read
Supabase for Everything
Auth, database, edge functions, realtime, storage. One service replacing five. Here's how I use Supabase as my entire backend.
Software Dev4 min read
My Code Review Checklist
What I look for when reviewing code: correctness, edge cases, naming, testing. Lessons from leading a team at Blinq.
Thoughts1 min read
AI won't replace engineers
It'll replace engineers who can't use AI. Different thing.
AI & LLM5 min read
AI Coding Tools Ranked: What I Actually Use
Claude Code vs Cursor vs Copilot vs Windsurf. Ranked by someone who uses them all for real work, not demos.
System Design8 min read
System Design: Distributed Cache
Consistent hashing, eviction policies, cache stampede prevention, and the Redis vs Memcached decision you'll actually face in production.
Tools3 min read
Local TTS at Zero Cost
Kokoro-82M: 337MB model, broadcast-quality speech, runs on a MacBook, costs nothing. Here's how to set it up.
Software Dev4 min read
Docker for Developers, Not Ops
Dev containers, multi-stage builds, compose for local dev. The Docker knowledge that actually matters when you're writing code, not managing infrastructure.
AI & LLM4 min read
Fine-Tuning Is Usually Wrong
A decision framework for when to fine-tune, when to prompt engineer, and when RAG is the right call. Most teams pick fine-tuning too early.
System Design8 min read
System Design: Payment System
Idempotency, double-entry bookkeeping, webhook handling, and PCI compliance — the system where bugs cost real money.
Product4 min read
Product Thinking vs. Coding
Why the best engineers think about users first. The skill gap companies don't test for.
AI & LLM5 min read
Claude vs GPT for Code Generation
An honest comparison from daily use. Where each model wins, where each fails, and what I actually reach for.
Thoughts1 min read
The best code is no code
Before writing a function, check if a library does it. Before a library, check if you need it at all.
System Design8 min read
System Design: Search Engine
Inverted indexes, query parsing, BM25 ranking, and autocomplete — designing an Elasticsearch-like search system from scratch.
Software Dev4 min read
Debugging Production Like a Detective
Log correlation, structured traces, and the debugging stories that taught me how to find bugs in production without adding console.log.
Tools4 min read
Automating Everything with Hooks
Claude Code hooks: auto-format, auto-test, auto-commit, auto-sync. My full setup for turning manual chores into background processes.
AI & LLM5 min read
Multi-Agent Orchestration Lessons
What I learned building a 23-agent system. When to split agents, when to merge them, and why most multi-agent architectures are over-engineered.
Product5 min read
Shipping Zero to One in 8 Weeks
Tactical breakdown of building AssessAI from nothing to 200+ source files, 630+ tests, and a working product. Architecture decisions and what I'd change.
System Design8 min read
System Design: News Feed
Fan-out on write vs read, ranking algorithms, and the caching strategy behind Facebook and Twitter's news feeds.
Software Dev4 min read
Testing Strategy for Solo Devs
When you're the only engineer, you can't test everything. Here's where to invest your testing time for maximum confidence.
AI & LLM5 min read
Building AI Agents That Don't Hallucinate
Grounding, tool use, and verification loops. What it took to ship agents that users actually trust.
System Design8 min read
System Design: Real-Time Chat System
Designing a WhatsApp/Slack-scale chat system — WebSockets, message ordering, presence indicators, and read receipts.
Thoughts1 min read
Most startups over-engineer
Postgres + Vercel + a single Next.js app covers 90% of use cases.
Software Dev4 min read
Next.js App Router Gotchas
Caching, revalidation, loading states, parallel routes. The things that bit me after 200+ source files on Next.js 15.
Tools5 min read
MCP Servers Explained
Model Context Protocol: what it is, how I use 9 servers daily, and a practical guide to connecting AI tools to everything.
AI & LLM4 min read
Prompt Engineering Is Dead
Structured outputs, function calling, and system-level constraints killed the artisanal prompt. Good riddance.
System Design7 min read
System Design: Notification System
Push, email, SMS — designing a multi-channel notification system with priority queues, delivery guarantees, and template engines.
Software Dev4 min read
React 19 Was Worth the Wait
Server components, the use() hook, Actions — what actually changed in practice after shipping a full product on React 19.
AI & LLM5 min read
Local LLMs Are Good Enough
Running Llama 3, Mistral, and Phi-3 on a MacBook. When cloud APIs are overkill and local inference makes more sense.
Tools4 min read
The Terminal Is My IDE
Claude Code as a primary development environment. 68 commands, tmux sessions, and why I rarely open VS Code anymore.
System Design7 min read
System Design: Distributed Rate Limiter
Token bucket, sliding window, and the Redis scripts that make distributed rate limiting actually work at scale.
Product4 min read
Why I Left My Job to Build AssessAI
The moment I decided, the problem I couldn't ignore, and what the first week looked like.
Thoughts1 min read
Hiring is broken
LeetCode interviewing in the AI era is cargo cult hiring.
AI & LLM4 min read
RAG Patterns That Actually Work
Vector search is the easy part. Here's what matters after you've embedded your documents.
Software Dev4 min read
TypeScript Patterns I Use Daily
Discriminated unions, branded types, const assertions, and satisfies. The patterns that actually make TypeScript worth it.
System Design6 min read
System Design: URL Shortener
Designing a bit.ly-scale URL shortener — Base62 encoding, redirect latency, analytics, and the caching layer that makes it fast.