In this post, I’ll talk in free form about my year at Etched. While I can’t talk closely about the projects I engaged in, I can speak instead about my growth as as a researcher and engineer.

When I started, I knew just about nothing. Well, almost nothing. I (mostly) solved one of their most important problems as a take-home interview, so I think I deserve some credit. Then I solved some (now seems simple) distributed systems problem. Having had no background in distributed systems, seems like it was sufficient.

What then?

Well, it was time to get to work. Learning computer architecture with no background in the field is strange. I’m not sure whether to approach it from the bottom up or top down. So far, I’ve attempted to do both, (aspirationally) reading a computer architecture textbook cover to cover, and studying our internal documentation and current papers on hardware-software codesign.

This was mostly an independent process, however - because I am more a “resident mathematician’’ of sorts at Etched, rather than a regular computer architect, the problems I work on can mostly be defined independently of the majority of the architecture. Call them “kernel problems”, because they are problems whose defining shape is by being really hard, but also isolated from the rest of the system.

After learning and project-doing for a whole year, I’m pretty proud of my progress. It has not been easy, both techincally (I’m reading a lot of papers now, and they are difficult to understand), and motivationally (when things don’t work, it’s hard to keep going). I’m also proud of my progress implementing my solutions. My training is in mathematics, not software engineering, so this is often considerably difficult. The last steps of any implementation seems to be the most difficult, because they are intellectually trivial, but can be painstakingly tedious.

What’s next?

I think my greatest worry as of now is that I am not publishing any work. My career seems like it is a standstill, although my growth is not. That’s why I would like to engage in efforts which can lead to publications. Both Stanford University and various research fellowships seem like they could be a great match at this point.

But first! I have a mathematics paper that I need to write. From my days studying geometric measure theory under Otis Chodosh, I seem to have a result which is deemed interesting by the geometric analysis community, which apparently solves a problem asked by Gromov.

That’s what next calls me - writing this paper, to perfect my practice of finishing things.

Stay tuned…