Capital One
AI Engineer 4 (AI Foundations: Benchmarking, Evaluation, and Explainability)
$197K–$246K
McLean, VA +3 · Full-time · Posted Oct 1, 2026
- Location
- McLean, VA
Cambridge, MA
San Jose, CA
New York, NY - Experience
- At least 4 years developing AI/ML algorithms or technologies with a Bachelor's degree, or at least 2 years with a Master's degree; at least 4 years programming. Preferred: 6+ years deploying AI solutions on cloud platforms.
- Education
- Bachelor's or Master's degree in Computer Science, AI, Electrical Engineering, Computer Engineering, or related fields.
- Visa sponsorship
- Will consider sponsoring a new qualified applicant for employment authorization.
About the job
Build and optimize foundational AI systems and infrastructure, including model evaluation, inference, agents, and responsible deployment for Capital One products.
What you'll do
- Design, develop, test, deploy, and support AI components such as foundation model training, LLM inference, agents, evaluation, and observability.
- Optimize production AI systems for scalability, cost, latency, throughput, and hardware utilization.
- Own architecture and technical direction for foundational AI systems, including reliability, governance, security, and compliance.
What they ask for
- Bachelor's degree in a related field plus at least 4 years developing AI/ML algorithms or technologies, or a related Master's degree plus at least 2 years of that experience.
- At least 4 years programming with Python, Go, Scala, CUDA, or Java.
Nice to have
- Experience leading AI system development and making cost, latency, throughput, and accuracy tradeoffs.
- 6+ years deploying scalable and responsible AI solutions on cloud platforms.
Pay
$197,300 - $225,100 USD base salary (Cambridge, MA); $197,300 - $225,100 USD base salary (McLean, VA); $215,200 - $245,600 USD base salary (New York, NY); $215,200 - $245,600 USD base salary (San Jose, CA)
Tools
Python, Go, Scala, CUDA, Java, AWS Ultraclusters, Huggingface, PyTorch
AI focus
Develop and optimize foundation models, LLM inference, agentic workflows, similarity search, guardrails, evaluation, governance, and observability for responsible, scalable AI systems.
Model-written summary · the original posting is authoritative