I've wanted a place to think out loud for a while. Not a feed, not a thread that disappears in a day, but somewhere ideas can sit, get revisited, contradict each other, and compound. So here it is.
Who's writing
I'm a product manager. Studied mechanical engineering (with a little materials science, Machine Learning, and technology entrepreneurship sprinkled in), started close to the metal in silicon and cloud hardware at Microsoft, spent years on Azure, and most recently was the founding PM at an AI infrastructure company building software for a multi-gigawatt compute buildout.
I’ve spent most of my career somewhere between hardware and models. Now I’m closer to the models themselves, working at a frontier lab on the part of AI that matters most to me.
What this is
Short notes, mostly. Thoughts too unfinished for a real announcement, too specific for a tweet, and too alive to disappear into my Notes app forever. Things I'm learning working in AI and product, the parts of the job that are harder than they look, and the occasional tangent.
I'd rather publish something half-formed and useful than wait for a perfect essay that never ships. Some posts will be technical, some will be about the work itself, and some will just be me figuring out what I think by writing it down.
What to expect
- AI and product. How the systems actually work, and the tradeoffs hiding underneath them
- The work itself. Working with engineers and researchers, translating ambiguity into plans, and prioritizing when it feels like the world is on fire
- Career, honestly. The non-linear path, the reps nobody sees, and what it feels like to reach past your grasp and have it work out
- Whatever I'm curious about
I'm not here to sell anything or perform expertise. I’m here because the best way I know to understand something is to say it out loud or write it down, over and over, until it’s clear. Not a thought leader, just unfortunately having thoughts.
No promises on cadence, only on honesty!
Open tabs
A running list of tools I'm loving, product thoughts I'm chewing on, and bets on where this is all going. It changes every post.
- Wispr Flow. Voice to text in any app. I dictate more than I type now, this post included.
- A question I keep circling: do the tools that win have the best UI, or just the least friction to start?
- Another one: when model behavior is the product, where does product management end and research begin?
- And a heavier one I’m carrying into the new job: how do you measure safety for the people least likely to show up clearly in your metrics?