Customer, operator, founder, builder
The short version: I've been the customer, the operator, the founder, and the builder. Product management is the job where all four get used at once.
Operator, founder, builder — one thread
I started as an operator. I ran two very different ecommerce operations: a US marketplace automation business — 77 client stores, a custom-built platform, a ~35-person ops team, $593K peak monthly GMV — and two private-label consumer brands on Amazon (A Baby Cherry and Decor & More), sourced from factories in China, Vietnam, and India and run with a team of eight. Operating at that scale, from building the platform to designing the profit splits, taught me the thing no PM course teaches: every product decision eventually lands on someone's P&L, usually mine.
Then I chased a pain I couldn't ignore. The worst recurring cost in those businesses was product photography. So I built QuicSnap — an AI-plus-humans photography service that made real revenue — and then Shutter Labs, the platform bet that a frontier model release eventually absorbed. That whole arc, including the ending, is the flagship case study; I'd rather you read the honest version than a summary.
Now I build AI products end to end. Since then I've shipped seven AI products — image platforms, ranking engines, agents — designing, building, and launching them solo, with the evals and decision logs public. Not because solo is the goal, but because it's the fastest way to keep my judgment calibrated on what models can actually do this quarter.
Why product, and why now
For years I ran everything — ops, growth, hiring, finance, product. The honest audit: product was the part I protected time for, the part I was best at, and the part I kept doing at 2am for fun. So this move is deliberate, not a landing pad between startups: I want to go deep on one product with a strong team around me, inside someone else's strategy, with a defined scope — and bring the operator's context to a PM seat instead of spreading it across nine jobs.
Two things I know about myself that matter to whoever hires me: I've spent years being the final decision-maker, which means I've learned the cost of every bad call the hard way — and I've also learned that the calls get better with more brains on them. The projects on this site are full of documented reversals for exactly that reason: I change my mind when the evidence says to, and I write it down when I do.
What I'm right for (and not)
0→1 and early products, AI-native teams, ambiguous problem spaces where the problem itself still needs choosing, products where model capability and product strategy have to be reasoned about together.
Pure optimization of a mature funnel, or roles where the PM never touches the model, the data, or the customer. You'd be paying for range I couldn't use.
The human bit
This block is reserved for 2–3 lines of genuine personality — hobbies, city, a strong opinion about something harmless. It's the only section that can't be researched.confirm
Say hello
The résumé is the compressed version of everything on this site — one page, no fluff.
LinkedIn confirm · GitHub confirm