AI Engineer · Technical Educator · Systems Programmer
I build production AI systems by day and teach the internals by night. After years of hitting walls because documentation stops where things get interesting, I started writing the articles and building the tutorials I wish had existed. No hand-waving, no "it's left as an exercise", no tutorials that only work in perfect conditions.
I've shipped LLM inference infrastructure, distributed training pipelines, low-latency trading systems, and a few open-source tools that somehow got traction. When I'm not coding, I'm probably reading about the same things but from a different angle.
Tools and domains I've shipped real things with — not just tutorials about.
Tutorials that teach you to copy-paste don't teach you anything. I try to explain why things work so you can reason through situations the tutorial didn't cover.
Toy examples don't prepare you for the real thing. Every piece of content here includes the failure modes, edge cases, and operational concerns you'll actually encounter.
Some things are genuinely hard. Pretending otherwise wastes your time. I'd rather say "this takes a week to internalise" than make you feel dumb for not getting it in five minutes.
The best engineers I know think across layers — from silicon to product. I try to connect the dots between ML math, systems code, and operational reality in the same piece.