Hannah Lee
Applied Scientist | ML Research to Production
Summary
Applied Scientist with 7 years bridging research and production ML. Authored 6 peer-reviewed papers with 900+ citations and shipped systems serving 20M users. Expert in deep learning, recommendation systems, and rigorous offline-to-online evaluation.
Professional Experience
Senior Applied Scientist at Atlas Research
Jan 2021 to PresentSeattle, WA
- Shipped a two-tower recommendation model serving 20M users, improving engagement by 11%.
- Led a research-to-production pipeline that reduced model iteration time by 60%.
- Designed offline metrics that correlate with online gains, cutting failed experiments by 40%.
- Mentored 3 interns, 2 of whom converted to full-time research scientists.
Applied Scientist at Quantum Analytics
Jun 2018 to Dec 2020Seattle, WA
- Developed a graph-based recommendation algorithm improving recall@20 by 14%.
- Published 3 papers at KDD, WWW, and RecSys; 400+ citations.
Technical Skills
Core: Deep Learning, Recommendation Systems, NLP, Probabilistic Models, Optimization, Tools: PyTorch, TensorFlow, JAX, Ray, Spark, Kubernetes, Languages: Python, C++, SQL, Scala
Projects
GNN4Rec | Graph Neural Network Recommender | Lead Author | github.com/hannahlee/gnn4rec
Jan 2022 to PresentOpen-source framework for graph-based recommendation research.
- 800+ stars; used in 15+ published research papers.
Education
Ph.D. in Computer Science (Machine Learning) at Stanford University
Sep 2014 to Jun 2018Stanford, CA
Selected Publications
Graph Neural Networks for Sequential Recommendation | KDD 2023 (oral, 8.2% acceptance)
2023Robust Offline Evaluation of Ranking Models | WWW 2022
2022