← home

AI consulting

I help teams ship AI systems that work

Most teams can get a demo running. The hard part is shipping something reliable: an agent that uses tools without going off the rails, a retrieval system that returns the right context, a model that runs where your data is allowed to live. That is the work I do with a small number of companies at a time, from early strategy through software in production.

I've spent over a decade building and scaling machine learning systems. I'm a Google Developer Expert in AI/ML, I created localGPT, and I teach these methods to a large community of engineers on YouTube. I bring that depth directly to your problem.

What I help with

Agentic systems
Designing agents that plan, use tools, and act, with the harness, evaluation, and guardrails that make them dependable instead of demo-only.
Retrieval and RAG
Grounding models in your own data with retrieval that goes beyond naive similarity search, tuned for accuracy on the questions that matter to you.
Local and open-weight models
Running capable models privately, on your own hardware or inside your cloud, for the cases where data cannot leave the building.
Strategy and enablement
Helping leadership and engineering cut through the hype, choose the right approach, and level up through workshops and hands-on guidance.

How I work

Advisory
Ongoing strategic guidance for founders and teams building with AI.
Hands-on builds
Architecting and building a system alongside your engineers.
Workshops
Focused, practical training to bring a team up to speed quickly.

Tell me what you're building.

The best place to start is a short email about your problem, your stack, and your timeline. I read every one.

Start a conversation →