Building and scaling backend data microservices that power StubHub’s end to end consumer experience

Developed and managed internal data services and backend processes to support the full stack of the BKLN platform, from account management to portfolio construction to trading

Developed novel fine-tuning pipeline for large language models used to power the thematic portfolio construction process on the BKLN platform

Implemented a novel machine learning pipeline for natural language processing tasks on data sources related to publicly traded companies with interpretable prediction explanations and visualizations that was put into production

Performed comprehensive research into data-intensive machine learning modeling approaches for other natural language processing tasks used by the BKLN platform

Interfaced with many Python libraries, including NumPy, pandas, SciPy, Matplotlib, scikit-learn, XGBoost, Keras, PyTorch, fastText, NLTK, spaCy, gensim, Google PEGASUS, ELI5

Awarded second place in the ACM Student Research Competition at SC19: The International Conference for High Performance Computing, Networking, Storage, and Analysis for a data-intensive cost/accuracy analysis of variant calling tools used in genomics and precision medicine research

Engineered machine learning models to predict the cost and accuracy for ensemble methods of variant callers with an accompanying trade-off graph, used to guide researchers’ decisions on the most cost-effective variant callers to use

Created an interactive Jupyter notebook with a pipeline for filtering data sets from the Genomic Data Commons, used by researchers at the National University of Singapore