At U.S. Bank, I build FastAPI backend services and React-powered AI applications, including a modular fraud investigation platform using LangChain and Python. I improved OpenSearch retrieval precision by 25%, maintained sub-50 ms p99 API latency, and increased AI content interaction accuracy by 20% through OpenAI integration.
Previously at Target, I developed recommendation pipelines using BERT embeddings and PySpark, increasing click-through rates by 7%, and built GPT-3.5 attribute-extraction tools that expanded catalog metadata coverage by 40%.
At Infosys, I delivered OCR, NLP, and secure Flask API solutions for regulated data workflows. I improved OCR accuracy by 35%, reduced data ingestion errors by 22%, and automated document extraction to reduce manual review by 40%.
