At Northern Trust, I built RAG financial research assistants for trust, fund administration, and investment policy retrieval. Their search relevance improved by 27%.
I also designed and deployed LLM inference services with p95 response latency under 420 ms. Distributed data pipelines I engineered support fraud monitoring across 1.8M+ daily records.
At Tata Consultancy Services (TCS), I fine-tuned transformer models for summarization, sentiment analysis, and entity extraction, increasing classification precision by 19%. I also developed predictive models that improved F1-score from 0.72 to 0.81.
On my BehaviorLedger AI project, I built a behavioral-finance system that detects behavioral drift and generates personalized financial guidance. I hold an M.S. in Computer Science from the University of Massachusetts, Boston.
