I indulged most of my childhood watching Japanese sci-fi anime like A Farewell to Weapons and Ghost in the Shell in Beijing, my hometown. When I came to the U.S. at the age of 15, I was attracted to the field of artificial intelligence on Kaggle. I was fortunate to study computer science at Carnegie Mellon University, where I learned rigorous problem-solving skills. I'm interested in mapping components of business decision-making processes to intelligent systems from the silicon up to complex software. I'm a member of the CMU automated learning systems group (Catalyst) and Neo. I'm currently working on Praxis, helping teams to integrate large, difficult external datasets.



I indulged most of my childhood watching Japanese sci-fi anime like A Farewell to Weapons and Ghost in the Shell in Beijing, my hometown. When I came to the U.S. at the age of 15, I was attracted to the field of artificial intelligence on Kaggle. I was fortunate to study computer science at Carnegie Mellon University, where I learned rigorous problem-solving skills. I'm interested in mapping components of business decision-making processes to intelligent systems from the silicon up to complex software. I'm a member of the CMU automated learning systems group (Catalyst) and Neo. I'm currently working on Praxis, helping teams to integrate large, difficult external datasets.
























responsibilities

  • machine learning & product @ Praxis Pioneering  - for one of the top record labels in the world, our team built and deployed 1. a large-scale asset valuation model for a wide range of entertainment IPs years & decades into the future and 2. a financial forecasting model for physical music (Vinyl, CD, Cassette) supply chain intelligence in an environment strained by COVID-19 and lacquer plants burning down
  • researching ML systems @ Catalyst Group, Carnegie Mellon University - we created a method that mixes optimized vendor libraries (e.g., cuDNN, cuBLAS, DNNL) for optimal performance in running complex neural networks. Our work, Collage, was published in PACT 2022 and adopted by OctoML into the codebase of their flagship project, TVM. 

projects


games

    Some of my personal favorites include:
  • Arknights: an exceptional narrative-based game with a unique East-Asian aesthetic. It is, in my opinion, one of the finest examples of its genre in the last decade.
  • Elite Dangerous: been playing for seven years and recently returned to it. The graphics are stunning, and the pace is slow enough for me to fully appreciate the view outside the cockpit, even when I am feeling stressed from work. I am currently working on acquiring and maintaining a Fleet Carrier in a location near a Low-Temperature Diamond hotspot to aid in my long-term mining endeavors. If you are interested, please reach out!
  • Civilization VI: a great game to kill time on a plane and ridiculously fun if you play it with a couple of friends you’ve known for a long time.

I’m mostly LA-based, while occasionally visiting Boston/NYC/Bay Area.