AI Engineer at Bank of America

Engineered deep learning models (TensorFlow, YOLOv8, MobileNetV2) for document classification and fraud detection; improved recognition accuracy by 35% and reduced manual review. Fine-tuned LLMs (LLaMA, Mistral, BERT, GPT-4) to automate client communications analysis and generate compliance summaries. Built a Claude-based generative AI chatbot to automate employee policy and account queries (10k+ monthly requests, 55% reduction in helpdesk workload). Implemented RAG pipelines with LangChain and LlamaIndex on Vertex AI to deliver contextual financial insights with 40% faster responses. Applied NER to regulatory filings to improve compliance traceability. Led distributed ML workflows in Databricks/PySpark for high-volume data; trained XGBoost models for credit risk. Integrated outputs into Tableau dashboards connected to PostgreSQL for real-time risk monitoring. Achieved $1.2M annual savings via fraud reduction and automation.