// what gets built here
const focus = [
“local-first AI systems”,
“on-device inference”,
“full-stack prototypes”,
];
iterate(focus, {
tests: 1547, passing: true
});
March 26, 2026 // 6 min read // Lessons & Reflections I have shipped over 30 projects across AI, mobile, enterprise, and open source. Each one taught me something the previous one could not. This blog is where I document those lessons. The name “Iterations” is deliberate. In machine learning, a model improves through iterations…
Explore by Topic
01
Prompt engineering, RAG architectures, fine-tuning workflows, and production LLM patterns.
8 posts
02
On-device inference, edge ML, privacy-first architectures, and offline-capable AI systems.
6 posts
03
End-to-end builds spanning React, gRPC, Python, Rust. From prototypes to production systems.
10 posts
04
Hard-won lessons, project retrospectives, and reflections on the craft of engineering.
5 posts
Latest
More posts coming soon.
About
I ship across AI, mobile, enterprise, and open source. Iterations is where I document the process — the architecture decisions, the debugging sessions at 2am, and the lessons that only come from building things end to end.
Every project here started as a question. How do you run LLMs on commodity hardware? How do you build an AI assistant that works offline? How do you ship a multi-platform app with one person? The answers are in the iterations.
Languages
Python, TypeScript, Rust, Dart, Swift
Shipped
30+ projects across AI, mobile, web
Tech Stack
AI / ML
PyTorch, Ollama, Whisper, Claude
Backend
Python, gRPC, PostgreSQL, Rust
Frontend
React, TypeScript, Vite, Tailwind
Mobile
React Native, Expo, Swift
Infra
NixOS, Docker, Supabase, GitHub CI