Joined Oxide when the whole company could still fit in a conference room. There's a lot that happens before the "main CPU" can even power on -- that's where I live. I worked out the strategy, chose the components, drafted and soldered the test boards, wrote the operating system (and the docs!), built the team, co-designed and implemented the platform secure boot architecture, debugged ONE MILLION I2C issues, and eventually shipped big racks of working servers with open-source firmware and no BIOS/UEFI.

I also try to act as a resource across the company on: difficult hardware/firmware debugging problems, component selection (particularly in MCUs), safe and unsafe Rust, ARM assembly language and program verification, etc.

I've been heavily involved in steering the company culture, starting before I joined (I had suggestions on the interview process and the terms of the employment agreement). We're imperfect but doing pretty decent.

Joined the Fuchsia project with the goal of making it the most secure operating system ever released. Turned out to be a poor fit.

Despite telling "other Bets" in Alphabet that they needed to obtain revenue as independent companies, Google saw our ML effort and simply absorbed it without X having a say. So, I was back at Google for a bit, doing basically the same stuff but with less optimism.

In fine Google tradition they subsequently cancelled the thing they had absorbed.

Joined one of the projects that grew out of Rapid Eval, building an inference-oriented machine learning acceleration chip for edge applications like robots and cars. Implemented both hardware/RTL and software in five languages (Haskell, Rust, Bluespec, Verilog, Tcl), managing people and projects, setting technical direction, and improving team culture.

The first (physically large and tremendously complex) chips off the manufacturing line worked, in part thanks to my verification efforts before we went to fab.

As an Alphabet company, X intended to both use this chip and sell it externally, but Google put the kibosh on both by absorbing us.

Took over management (with Phil Watson) of the team responsible for spinning up -- and shutting down -- potential "moonshot" projects. We improved culture and effectiveness, working directly with project teams, X leadership, and stakeholders across Alphabet. Rapid Eval started more projects under our watch than in its entire four-year history up to that point...none of which I can talk about yet.

I was promoted to Senior Staff Software Engineer during this time.

I appear as a side character in the Atlantic article linked below.