01 / 04
Start with how the work actually moves, not with the chatbot. Redesign the workflow around what language models are genuinely good at — then prove it with evaluations before it ships.
What I do
Most internal AI efforts stall because a chatbot gets bolted onto a process that was already broken. I start from the workflow itself — how work actually moves, where the handoffs cost you — then redesign around what language models are genuinely good at, and prove it with evaluations before it reaches production.
An independent read on what you have built. Architecture, model choices, data pipelines, roadmap feasibility, and team capability — assessed plainly, including the question most reviews avoid: whether the AI in the product is doing real work or decorating a demo.
Where AI investment should go, and where it shouldn’t. Which bets compound with what you already own — proprietary data, institutional knowledge, distribution — and which are expensive detours into somebody else’s advantage.
Technical reality translated into a story capital allocators can underwrite: what the technology does, why it is hard to copy, and how it converts into revenue — in language that survives contact with a technical diligence call.
Current work
Engagement in progress
I run most of the product discovery and definition for an early-stage team — customer research, problem framing, scope, and the written definition engineering builds from. The founders hold the vision; my job is turning it into decisions the team can act on week to week, and keeping the roadmap honest about what is actually validated.
Engagement in progress
Full product definition and hands-on development for a second startup — defining the product and building the working demos that carry investor conversations. Close enough to the code to ship it myself, which means the story in the room and the thing on the screen stay the same story.
Client names withheld while engagements are active. Named references available on request.
Engagement models
Fixed-fee · Project
A focused 2–3 week engagement to assess your current AI and product posture, identify gaps, and deliver a prioritized action plan. The cleanest way to start before a longer commitment.
Start a sprint →Retainer · Ongoing
Embedded fractional CPO engagement at 2–3 days per week. Product strategy, roadmap leadership, team coaching, and direct engineering partnership across a minimum 3-month initial term.
Discuss an engagement →Advisory · Lightweight
Monthly strategy sessions plus async access. Ideal for founders and executive teams who need a thinking partner — not execution bandwidth — on product and AI decisions.
Explore advisory →