AI consulting for systems that work in practice
I help teams move from an AI idea to a useful, testable system. That can mean clarifying the use case, designing an agentic workflow, building an initial implementation, or improving the way an existing AI system is deployed and evaluated.
Discuss an AI projectTell me the problem, the team, and what you have tried so far.
When to bring me in
- You have a promising AI use case but need to decide what should be automated, what should stay with people, and how to measure success.
- A prototype works in a demo but needs better context, integration, evaluation, or operational controls before people can rely on it.
- Your team needs an engineer who can move between implementation, architecture, and technical strategy while explaining the tradeoffs clearly.
How I approach the work
Frame the problem
Map the workflow, constraints, data, and failure modes. Choose a narrow first outcome that can be tested.
Build and connect
Implement the workflow with the right tools, integrations, context, and human checkpoints.
Evaluate and improve
Define useful tests and feedback loops, then decide what is ready to run and what needs more evidence.
The work behind the approach
I founded Tabula and maintain ctx, an open-source tool for grounding coding agents in a queryable model of a codebase. My earlier engineering work spans SAS, Ericsson, IKEA, H&M, and Tele2.