AI Engineer 4 (Gen AI Platform Services)

Capital One is creating responsible and reliable AI systems to transform banking through machine learning and scalable AI infrastructure. The AI Engineer 4 will build, optimize, deploy, and support foundational AI systems, including large language model inference, agentic workflows, model evaluation, governance, and observability. The role also owns complex AI architectures, defines reliability objectives, leads technical reviews, and mentors engineering associates.

Responsibilities:

• Partner with a cross-functional team of engineers, research scientists, technical program managers, and product managers to deliver AI-powered products that change how our associates work and how our customers interact with Capital One
• Design, develop, test, deploy, and support AI software components including foundation model training, large language model inference, agents and multi-agent workflows, similarity search, guardrails, model evaluation, experimentation, governance, and observability, etc
• Leverage a broad stack of Open Source and SaaS AI technologies such as AWS Ultraclusters, Huggingface, VectorDBs, PyTorch, and more
• Invent and introduce state-of-the-art foundation model optimization techniques to improve the performance — scalability, cost, latency, throughput — of large scale production AI systems
• Contribute to the technical vision and the long term roadmap of foundational AI systems at Capital One
• Own the end-to-end architecture for complex AI systems - ensuring maintainability, observability, and ethical alignment
• Define and maintain service-level objectives (SLOs) for AI reliability, including latency, uptime, and model performance drift
• Collaborate with infrastructure engineering to optimize GPU/TPU utilization and accelerate model inference pipelines
• Lead cross-functional technical reviews for new AI system deployments, ensuring security, data governance, and compliance standards are met
• Mentor Principal and Senior Associates on scalable design, performance tuning and research-to-production translation