Services

AI Engineering Consulting for Small Teams That Ship

AI engineering consulting for founders and small engineering teams. Set up AI coding assistants across your repositories, add the guard rails that keep them safe in production, and connect AI agents to the systems your business runs on.

I run a hardware and software company with AI doing a large share of the engineering: firmware, backend, web, mobile and the internal tools around them. It works because of how the assistant is set up, not because of which assistant it is.

I have also seen what happens when it is set up badly: a push to the wrong repository, a deploy broken at one in the morning, a cost-saving change that quietly made the work worse. AI engineering consulting with me means getting the benefit of those mistakes without making them yourself.

What teams come to me with

The assistant is fast inside one repo and lost between them

A feature that touches the backend, the web app and the mobile app turns into three separate conversations. I help set up one workspace where the assistant sees how the pieces connect.

It commits, pushes or deploys when it should not

Written instructions are not enough for anything that reaches production. I help put hard gates in place, so approval is required by the tooling and does not depend on memory.

The bill grows faster than the output

Most of the cost is rarely where people assume. I measure where it goes and separate the savings that are free from the ones that make the work worse.

Instructions and memory that have gone stale

A long instruction file that nobody maintains misleads the assistant on every task. I help decide what belongs in it, what belongs next to the code, and what should be deleted.

Wanting an agent that does real work

An agent that reads your CRM, answers on a phone line or acts on your systems is mostly a wiring problem. I have built these and know which parts are hard.

Not knowing what to trust it with

Some work can be handed over completely and some should never be. I help draw that line for your product and your team.

Who gets the most from this

  • Founders building alone or with a few engineers who want to cover more ground.
  • Engineering leads introducing AI assistants to a team and wanting rules that hold.
  • Companies with several repositories where AI help stops at each repository's edge.

How we can work together

A working session
One focused call on one problem. I read your material beforehand so the time goes to the problem.
A written review
I go through your design, code or plan and send back findings in order of priority, with what I would change and why.
Ongoing advisory
Regular calls while you build, with questions answered in between. Suited to a team working through a hard stretch.

See how I work

Questions people ask first

Which tools do you work with?

I work with Claude Code every day across more than ten repositories, and with the Claude API for agents. The principles carry over to other assistants.

Do you write our code for us?

I set up how your team works with the assistant and review the results. The code stays yours, written in your repositories by your people.

Can we see how you work before talking?

Yes. The guides and articles on this site show the setup I use, with the files to download.

Talk to me about AI engineering consulting

Tell me what you are building and where it is stuck. I read every message myself.

Message me on LinkedIn