Welcome! You’ve gradient descended into my local minima. I’m Han, and I specialize in building machine learning systems.

I write about systems I have built: production AI, search and recommendation, evaluation, and the engineering and organizational failures around them.

I currently serve as Director - Machine Learning at Moody’s. I lead teams that build and operate custom LLMs, generative AI applications, and search and discovery systems for financial data.

Previously, I led data science engineering at WalletHub. The recommendation system my team designed increased click-through rates by 10% and conversion rates by 40%. I also designed an MLOps platform with Airflow, AWS Lambda, Bitbucket Pipelines, SageMaker, MLflow, Terraform, and CI/CD.

Earlier, I worked at Workhuman as an NLP Engineer, AMD as an IC Design Engineer, and Ericsson as a Hardware Engineer.

Side Quests

In the metaverse, I co-founded Eminence (Blackrock), a World Top 6 PvE guild in World of Warcraft (Classic + TBC) back in 2004. We operated remotely across several continents and time zones, coordinating multiple raid teams of over 120 full-time players. We were 15 years ahead of the broader transition to remote work.

Investing

I am the founder and portfolio manager of Calabazas Creek Advisors.

Previously, I worked on the sell side as Portfolio Co-Manager of the Firsthand Technology Opportunity Fund and on the buy side as an Equity Research Associate at Global Crown Capital.

During my investment career, I appeared on the cover of Barron’s.

Background

I hold a master’s degree from San Jose State University and a bachelor’s degree in Electrical Engineering and Computer Science from University of California, Berkeley. My research interests span machine learning systems design, information retrieval, and recommendation systems.

Contact

Han Lee
San Francisco Bay Area, California, United States
Email: lee [dot] hanchung [at] gmail [dot] com