Product manager who builds things.
UX discipline plus AI in the loop equals days where it used to take quarters. I take fuzzy problems, find the real job underneath, and ship the smallest version that solves it. Real working prototype before the kickoff meeting ends, evals running in the background.
- Fractional PM at an AI-for-VC due-diligence startup
- Looking for senior or staff PM seats at AI / product-led teams
- Toronto / remote
mySommelier — Five LCBO Vintages bottles you'd love this Saturday, picked by an agent that updates its taste model every time you rate a wine. Live at my-sommelier-rinalabs.vercel.app. Ships with three eval layers (data quality, retrieval NDCG@5, append-only trace logging). Provider-agnostic, so swapping Gemini for OpenAI is a config change. Case study.
Sketch. Prototype. Put it in front of a user. Adjust. The bottleneck moved from "can we build it" to "is it the right thing," and that's a UX question, which is what I trained on.
When the team is stuck, the answer is almost always go talk to a user this afternoon. Not next sprint. Today.
When an AI feature looks reasonable in a screenshot but might be confidently wrong, I write the eval before I trust the model.