Data Scientist. Researcher at LBNL. Student at UC Berkeley MIDS. Columbia University alum.
LinkedIn · San Francisco, CA
I work on applied machine learning and data science, spanning recommender systems, forecasting, LLM evaluation, NLP, and causal experimentation. Below are selected projects from coursework and independent work.
| Project | Description |
|---|---|
| datasci241-llm-prompt-consistency-experiment | Controlled experiment (with power analysis and treatment/control design) testing whether prompt-engineering interventions improve LLM response consistency. Consists of A/B testing methodology applied to LLM evaluation. |
| datasci207-movie-recommender-matrix-factorization-cf | Movie recommender system comparing matrix factorization (SVD) and item-item collaborative filtering on MovieLens 1M |
| datasci203-firm-characteristics-stock-price-event-study | Event study on associations between firm-level characteristics and short-term stock price changes |
| ashrae_competition | Comparison of neural network architectures (dense NN, entity embeddings, CNN-1D) for time series forecasting of building energy consumption in the ASHRAE Kaggle competition (1,000+ buildings) |
| Spatial-Wind-Power-Forecasting | Spatiotemporal forecasting of wind power output using ARIMA modeling |
| ai4doc | Pipeline for collecting, filtering, and annotating 22K+ SEC regulatory filings for document layout detection and OCR |
Python, R, SQL, PyTorch/scikit-learn, causal inference & experiment design, time series forecasting, recommender systems, NLP/LLMs