Peter Tsai Email & Phone Number

San Ramon, CA, US

Chicago, IL, US

Redwood City, California, US

Assisted major players across the materials and chemicals industry build ML models from their data to accelerate R&D projects in an adaptive learning workflow - using uncertainty quantification to balance exploration against exploitation.Assisted subject matter experts in designing custom input features derived from their domain knowledge.Engaged with.

St. Gallen, CH

Contributed to significant improvement of existing classification models by strategically applying iterations of error analysis, feature engineering, rectifying dataset imbalance, and hyperparameter tuning.Made predictive models more interpretable by refactoring existing model architectures, and incorporating engineered features from subject matter.

San Francisco, CA, US

Built NBATicketSavant, an app to predict the monetary risk of waiting on purchasing a ticket to a professional basketball game.Assembled a time-series SQL database by daily scraping more than 200,000 ticket listings via StubHub’s public APIs.Applied logistic regression, support vector machine, and random forest algorithms on data to predict monetary risk.

St. Louis, MO, US

Applied K-means clustering algorithm to identify and classify distinct features in complex metallic microstructures, resulting in ~400% improvement in accuracy over conventional binning techniques.Carried out a comprehensive statistical study of microscale mechanical properties in a metallic glass, resulting in an improved understanding of property.

St. Louis, MO, US

Developed a novel combinatorial synthesis methodology with additive manufacturing equipment to accelerate the discovery of high-performance metallic alloys, reducing exploration time by ~200x compared to conventional serial methods.Implemented an automated nanoindentation strategy to rapidly screen materials from combinatorial libraries that can contain.

Medium Article: A Metallurgical Scientist's Approach to Predicting NBA Team SuccessUsed Python and its data scraping modules to extract and reconstruct shot chart data for all 400+ currently active players in the NBA.Applied clustering algorithms to quantitatively define categories of players based on their shooting tendencies. Discovered significant.

Columbus, Ohio, US

Research focused on developing a combinatorial approach for designing bulk metallic glass alloys. The primary equipment used was an Optomec 750 LENS 3d printer for metals.

Doctor Of Philosophy (Phd), Materials Science And Engineering

Master Of Science (Ms), Materials Science And Engineering

B.S., Materials Science And Engineering