San Francisco Bay Area
- Engineered a RAG-powered research assistant integrating LLMs with retrieval pipelines to automate charting and research note generation, boosting analyst productivity and enabling faster response to market-moving events. - Built automated data pipelines in Python to ingest, transform, and validate performance metrics across 20+ strategies, cutting manual reporting time by 60%. - Developed Markov regime-switching models and LASSO regression systems to detect market regimes and predict commodity returns, informing portfolio positioning decisions on $10M in assets.
- Developed algorithms and strategies to predict securities performance and risk - Developed data pipelines to collect and parse text data from earning call transcripts using Python. - Applied Natural Language Processing (sentiment extraction, topic modeling) to transcripts data.
- Developed wearable tech to help users self-learn dancing - Created motion representations of tap dance moves and their sequences - Designed and implemented sensors and algorithms to capture, encode and analyze the movement data, and actuators to convert it back to physical prompts.