AI/ML Engineer – Python at Capital One

Built Python-based data pipelines using Pandas and SQL to clean, transform, and feature-engineer financial log data, enabling exploratory analysis that identified early bias signals during generative AI safety experiments and supported large-context LLM inputs. • Integrated LLM APIs using Python with focused prompt engineering and retrieval-augmented generation patterns via LangChain to enhance cybersecurity simulations, improving response accuracy while minimizing unnecessary data exposure. • Designed and evaluated basic machine learning pipelines using scikit-learn for classification and regression use cases, validating model performance with standard metrics and running lightweight PyTorch experiments to test multi-agent consistency and reliability. • Developed and exposed backend functionality through RESTful APIs using FastAPI and Flask, managing JSON data flows and persisting benchmark results in MongoDB, with authentication, logging, and error handling to support scalable research workflows. • Applied modern software engineering practices by using GitHub for version control, writing pytest unit tests, and containerizing services with Docker, enabling CI/CD pipelines through GitHub Actions and supporting collaborative, Agile development across research teams.