Independent AI Engineering Consultant
I help established companies ship LLM-powered features that hold up under real users — without the research-lab detours. RAG systems, agent workflows, evaluation harnesses, and the boring infrastructure that makes them reliable.
LLM application development is the core practice. The rest exist to support it.
Custom LLM features end-to-end — built with Claude, GPT, Gemini, or open-source models depending on what fits the problem. From first prototype to a service your team can operate.
Evaluation harnesses, regression tracking, and quality measurement for LLM features. Catches the failure modes that don’t show up in a demo but find every user on day one.
Advisory work on what to build, which models fit, build-vs-buy, and how to sequence an AI roadmap. Useful before you commit engineering budget to a direction.
Latency, cost, observability, and safety for LLM features already in flight. The work that turns a working prototype into a service the on-call rotation can sleep through.
I build with existing foundation models — I’m not a model-training shop. If you need ML research or custom model training, I can point you toward people who do that work.
Predictable scope, predictable price, predictable handoff.
Thirty minutes. We talk through what you’re trying to ship, what’s in the way, and whether I’m the right fit. No charge, no pitch.
A written scope with deliverables, timeline, and a fixed price — or a clear retainer for advisory work. You see it before you commit.
I work as an embedded engineer with your team — typically four to twelve weeks per engagement. Weekly demos, async-first, your code stays in your repo.
Documentation, runbooks, and the context your team needs to operate what we built. Available for follow-on work if you want it.
I’m a software engineer who has spent the last several years focused on building applications with large language models. I work best with teams that already have a product and real users, and want to add LLM capabilities without slowing the rest of the roadmap.
I take on a small number of engagements at a time so each one gets full attention. Remote-first, US-based, comfortable with whatever cloud, language, or framework you’re already running.
The fastest way to start is an email describing what you’re working on and what you’d like help with. I read every one and reply within a couple of business days.