senior pm · ad platforms and ai
I own platform businesses and build on them. These are the calls I made, including the ones I would make differently.
Work
-
01 / supply
Rebuilding how publishers target their inventory
Publishers had to describe their inventory placement by placement, which meant maintaining thousands of configurations to express something simple. The cheaper option was to leave it alone and let them absorb the work. I argued for rebuilding the targeting model instead.
Lesson
Treat the customer experience of a transition as part of the design, not as something that follows it.
- over a year
- staged rollout
-
02 / platform
Testing a fix before we built it
Two engineering leaders proposed a solution to a targeting gap and were confident it would hold. I doubted it on a hunch, spent a weekend building the analysis to find out, and it turned out the design would have left 14% of revenue exposed.
Lesson
When repeated fixes keep leaving edge cases behind, the problem is not understood yet.
- one weekend
- 14% revenue exposed
-
03 / personal
The spreadsheet that became a toolkit
I set out to rebalance our portfolio in a spreadsheet, realized how tedious it would be to maintain, and decided to find out what I could build instead. It turned into a suite of tools, and a lesson about scope.
Lesson
When building feels cheap, the discipline that protects you is the one you stop noticing you dropped.
- ongoing
- python
About
I work on the platforms that sit between publishers and the demand that pays for their content. My career has moved steadily toward the decisions behind technical products, and I got there by taking on that work before it was mine to do. Lately that includes writing code with AI agents.
In January 2026 I went from AI skeptic to AI native, now building work and personal tools daily.