WritingI'll flesh this out more later. For now, here are some links, in roughly chronological order:
- Cutting over large systems was a post I wrote for the Affirm tech blog. I spent the last three months or so of 2018 designing and rolling out a top-to-bottom rewrite of our promotional messages stack, from database to rendering to API endpoint. The short version is this: we needed to roll out a massive thing with (a) zero downtime and (b) zero regressions. Our solution: run both systems in parallel and verify (asynchronously) that the result was the same. It worked beautifully.
- Preparing Texas Students for the Cyber Economy was a report I cowrote with Dr. Carol Fletcher (now at TACC) on the state of computer science education in Texas high schools. Over the course of six months, we assembled a group of high school teachers and administrators, policy experts, industry professionals, and computer science professors to figure out how to improve funding for CS education. The report was submitted to the Texas Legislature, which quickly followed our recommendations and increased funding (by around $7,000 per course offered).
- An Artificial Intelligence Approach to Redistricting was my undergraduate thesis. The United States has a really strange system of electing representatives, based on district maps drawn by partisan state legislatures once every ten years. It's incredibly easy to manipulate district maps to shift a majority one way or another, and it happens all the time. My approach used a newly-devised method of measuring partisan skew, the efficiency gap, to score randomly-generated district maps. I used a genetic algorithm to combine those district maps and extract the best, least-skewed ones, generating a fairer mapping.
- This op-ed was an essay I wrote for the Daily Texan, the UT student newspaper. UT Austin is one of a very small number of schools that place their computer science department in the College of Natural Sciences. This creates a system where the CS department, which is far and away the largest money-maker for the College, funds much smaller departments. While in general I don't really mind that, the CS department is itself critically underfunded; core classes required for graduation are often filled up before everyone who needs to take them can get a seat. I proposed that UT move computer science and a few other related departments to their own college, which would give the department more control over funding.