Hi, I'm Hamel

I'm an ML engineer and independent consultant at Parlance Labs. I spend most of my time helping teams build AI products. Previously, I did applied ML at GitHub and Airbnb.

What I'm working on

I'm working to bring data science back to AI: helping teams debug, analyze, and measure their systems. I call this "evals," and after doing it across 35+ AI products, I co-authored Evals for AI Engineers (O'Reilly), covering error analysis, LLM-as-a-judge, synthetic data, production monitoring, and building data flywheels. I also co-teach a course on evals on Maven.

I write about what I learn at hamel.dev. Some recent posts:

Date	Post
Jan 2026	LLM Evals: Everything You Need to Know
Jul 2025	Stop Saying RAG Is Dead
Mar 2025	A Field Guide to Rapidly Improving AI Products
Dec 2024	nbsanity - Share Notebooks as Polished Web Pages in Seconds
Nov 2024	Building an Audience Through Technical Writing: Strategies and Mistakes
Oct 2024	Using LLM-as-a-Judge For Evaluation: A Complete Guide
Oct 2024	Concurrency Foundations For FastHTML
Jul 2024	An Open Course on LLMs, Led by Practitioners
Jun 2024	What We've Learned From A Year of Building with LLMs
Apr 2024	Debugging AI With Adversarial Validation

Open source

I've contributed to tools across ML infrastructure, developer experience, and data science workflows: machine learning frameworks, workflow orchestration, Jupyter tooling, and code search. Full list here.