What Topics Does Hamel Husain Write About on hamel.dev?

Hamel's Blog (hamel.dev) focuses on applied AI engineering, machine learning, and data science, with a strong emphasis on evals—the practice of debugging, analyzing, and measuring AI systems. The blog is best suited for engineers, product managers, and data scientists who build or evaluate AI products and want practical, experience-based guidance rather than general AI commentary. Posts are long-form and drawn from the author's work at Airbnb and GitHub, including early LLM research used by OpenAI for code understanding.

Core Topic: AI Evals

Evals are the blog's central theme. Hamel describes his current work as "bringing data science back to AI"—helping teams debug, analyze, and measure their systems. Related posts include:

• AI Evals: Everything You Need to Know (9/18/26) — a broad overview of the discipline.
• How To Use Decision Models for Evals (10/20/26) — applying decision models to evaluation design.
• Do Automated Evals Work? (7/11/26) — examining the limits of automated evaluation.
• Using LLM-as-a-Judge For Evaluation: A Complete Guide (10/29/24) — a complete guide to LLM-based judging.
• Selecting The Right AI Evals Tool (10/1/25) — comparing evaluation tooling.
• Inspect AI, An OSS Python Library For LLM Evals (6/23/25) — an open-source Python library for evals.
• Evals Skills for Coding Agents (3/2/26) — evaluation skills in the context of coding agents.
• Claude's new auto eval tool (9/30/26) — commentary on a new automated eval tool.

If you are new to the concept, the AI Evals Guide linked from the site is the recommended starting point for common questions.

AI Product Engineering

A second cluster covers building and improving AI products in practice:

• AI Product Engineering Notes (8/12/26)
• A Field Guide to Rapidly Improving AI Products (3/24/25)
• "It's Hard to Eval" Is a Product Smell (6/29/26) — argues that evaluation difficulty often signals a product problem.
• Your AI Product Needs Evals (3/29/24)
• What We've Learned From A Year of Building with LLMs (6/1/24)
• Thoughts On A Month With Devin (1/19/25) — firsthand observations from using an AI coding agent.
• Stop Saying RAG Is Dead (7/11/25)

LLM Infrastructure and Engineering

The blog also covers lower-level LLM engineering concerns:

• Tokenization Gotchas (12/17/23)
• Optimizing LLM latency (10/15/23)
• vLLM & Large Models (10/28/23)
• Tools for curating LLM Data (11/15/23)
• Is Fine-Tuning Still Valuable? (3/27/24)
• Concurrency Foundations For FastHTML (10/10/24)
• Debugging AI With Adversarial Validation (4/12/24)

Data Science, Tooling, and Technical Writing

Beyond AI, Hamel writes about the data science craft and the tools he uses:

• The Revenge of the Data Scientist (3/26/26)
• Why I Stopped Using nbdev (1/18/26) and On commercializing nbdev (5/30/23)
• nbsanity - Share Notebooks as Polished Web Pages in Seconds (12/13/24)
• Notebooks in production with Metaflow (2/9/22)
• ghapi, a new third-party Python client for the GitHub API (12/18/20)
• Dokku: my favorite personal serverless platform (1/9/24)
• Why Should ML Engineers Learn Kubernetes? (1/16/23)
• Building an Audience Through Technical Writing: Strategies and Mistakes (11/30/24)

Learning Resources and Following Along

The site lists two structured ways to learn from Hamel:

You can also subscribe to the newsletter to receive new long-form posts, and the site includes a "Follow Me" section for ongoing updates.

How to Decide If It's Worth Subscribing

• Subscribe if you work on AI products and need concrete methods for evaluation, debugging, and measurement—especially if you've found "it's hard to eval" to be a recurring blocker.
• Start with the AI Evals Guide if you want the conceptual foundation before reading individual posts.
• Browse the archive by date if you're interested in a specific infrastructure topic (tokenization, latency, vLLM) or in the data science and tooling essays.
• Consider the course only if you want cohort-based instruction; the blog and guide cover the same subject matter in written form.