Joyce Xu
AI/ML researcher specializing in NLP, reinforcement learning, and distributed computing.
Aspiring bartender, DJ, and unemployed white male p'dcaster.
Chaotic neutral.
AI/ML researcher specializing in NLP, reinforcement learning, and distributed computing.
Aspiring bartender, DJ, and unemployed white male p'dcaster.
Chaotic neutral.
Hello! I am an undergraduate studying computer science and urban studies at Stanford. These days, I mostly spend my time thinking about housing and cities, and behind them the seemingly insurmountable racial and economic injustice that dominates our arc of history and the inherent challenges of "optimizing" complex systems that are rooted in the human. But, for the purposes of biding time while I think about tackling these at a larger scale, let's pretend this website is a self-evangelizing tech blog like any other.
I've had the incredible fortune of working with some truly brilliant people these past few years. I am currently working on data infrastructure for large-scale urban computing at Replica. I've pursued natural language processing research under Chris Manning at the Stanford AI Lab, and published at ICML/CVPR on adversarial robustness with the support of UnifyID's AI fellowship. Previously, I collaborated with some amazing researchers and engineers on multi-agent reinforcement learning at DeepMind, real-time object detection at the AI consultancy ThinkTopic, and knowledge extraction and inference at the Zuckerberg-backed personalized learning startup Volley. (If you're here to read about something other than my self-evangelization, please look into our work at the Anti-Eviction Mapping Project).
I began my journey self-taught in CS/AI, so I feel somewhat strongly about accessibility of code + research. At one of my previous jobs, we developed and open-sourced our own ML framework in Clojure called Cortex, on which I've given a few talks. Pieces of the reinforcement learning libraries I extended at DeepMind are now being open-sourced. You can also find me writing for Hacker News, KDNuggets, AITopics (a publication of the AAAI), Towards Data Science, and Startup Grind. Finally, there were a few naive months when I believed "algorithmic fairness" would mean anything in the vacuum of other change, but from it fortunately I produced a (fairly ok) presentation relating housing justice to ML research in bias.
When I am not busy procrastinating, I am busy liberally distributing this obnoxious tech salary. If I am home in Colorado, come find me skiing; if I am in the bay, come say hello at the next rave at Bill Graham. At 4am most nights, I might be tinkering with music production and I might be reading New Yorker articles on my bed, but as long as you can handle my sleep schedule, feel free to get in touch. I am always happy to chat - particularly if you are also thinking about housing and cities.