OPERATOR-ADVISOR · PRODUCTION AGENTIC AI
I'm a product leader who's been building inside the AI shift for a decade. Now I help others build through it.
Recent essays
15+ years shipping product at
You've been in pilot mode for 18 months. Pilot purgatory has a cost — and it's not just time. Your board isn't asking the wrong questions. They're asking the right ones, and no one can answer them yet.
Most AI initiatives fail between the strategy slide and the production deploy. The gap isn't strategy. It's the person who can turn the strategy into a working system.
What I do
Get unstuck. Diagnose, prioritize, operate alongside leadership.
Pressure-test the bet before you stake the company on it.
Stop reinventing the substrate. Your engineers will thank you.
Field notes
Series B fintech · Q4 2025
Problem
Inference costs tripled month-over-month. Engineering couldn't trace which model calls were driving the spike. Board was asking CFO questions the team couldn't answer.
What shipped
Cost routing layer in 6 weeks — structured model-selection logic, full audit trail mapping every dollar of inference cost to a feature call.
Outcome
Inference cost reduced 58% while throughput increased. First time the team could answer a board question about cost-per-transaction.
Consumer marketplace · Q1 2026
Problem
Two years of AI feature work, one agent in production, zero eval framework. Releases were manual spot-checks by a PM with a spreadsheet.
What shipped
Eval harness against 3 months of production logs — 200 labeled examples, automated regression suite, a deploy gate that blocks on confidence drop.
Outcome
Eval cycle dropped from 2 weeks manual to 3 hours automated. Next model upgrade shipped in one sprint instead of a quarter.
PE portfolio diligence · Q3 2025
Problem
Operating partner needed AI maturity read across 3 portcos before writing a value creation plan. No internal AI expertise to separate real from pitch-deck.
What shipped
48-hour diligence protocol per portco — architecture review, production log analysis, team capability assessment, gap-to-value report.
Outcome
One portco had real AI and a credible roadmap. Two had $600K of technical debt to unwind. The value creation plan was built on reality, not the pitch.
Running a similar diligence process? The cheat sheet is here →
Find your fit
If you lead AI inside a company
You're on the hook for pilots graduating to production. Read what that work looks like when someone else has already made the mistakes.
If you're a founder building an AI product
You're making architecture calls that feel irreversible. Read how to pressure-test the bet before you spend a quarter on the wrong one.
If you're an investor working with portfolio companies
You need a read on where AI capability is real and where it's pitch-deck. Read what that diligence looks like when someone can read the logs.
Selective by necessity — I work with a few teams at a time. We figure out together if there's real fit.