AI/ML Engineer at Capital One

Led retrieval-augmented generation initiatives using Azure OpenAI, LangChain, Pinecone, and enterprise document repositories to boost knowledge retrieval accuracy by 43% and cut research time by 58%. Automated 65% of recurring customer and operational requests with LLM-based applications, improving consistency and productivity across multiple business teams. Delivered ML services for fraud analytics and customer intelligence using Python, PyTorch, FastAPI, and Kubernetes, shortening deployment timelines by 47%. Implemented end-to-end MLOps pipelines (MLflow, Airflow, Docker, GitHub Actions) to improve reproducibility and monitoring, achieving 45% deployment efficiency gains. Improved fraud detection rates across millions of transactions using PyTorch and XGBoost, and built real-time feature pipelines on Spark and Kafka to reduce latency by 38%.