JPMorgan Chase
Applied AI ML Director - AI Agents and Agentic Systems
$223K–$325K
Palo Alto, CA +2 · Director · Full-time · Posted Sep 18, 2026
- Location
- Palo Alto, CA
Seattle, WA
New York, NY - Experience
- 10+ years applied experience; 10+ years building large-scale machine learning systems and platform services.
- Education
- Formal training or certification on applied artificial intelligence and machine learning concepts; advanced degree in Computer Science, Machine Learning, or a related field preferred.
About the job
Build and evolve an enterprise Agent SDK and agentic AI platform, turning emerging techniques into reliable, evaluated, production-ready capabilities.
What you'll do
- Architect and implement Agent SDK capabilities, reference implementations, specialized agents, and reusable components.
- Design evaluation methods for agent behavior and optimize agentic systems for performance, accuracy, reliability, latency, and cost.
- Partner across product, data, platform, and application teams to guide technical decisions and drive adoption against business outcomes.
What they ask for
- Formal training or certification in applied AI and machine learning concepts.
- 10+ years of applied experience and 10+ years building large-scale ML systems and platform services used by multiple teams.
- Strong software engineering skills, including end-to-end delivery; extensive experience with PyTorch or TensorFlow.
- Hands-on agentic and generative AI design experience, plus cloud and Kubernetes production workloads and high-performance ML systems.
Nice to have
- Experience contributing to or optimizing open-source ML frameworks or platform tooling.
- Advanced degree in Computer Science, Machine Learning, or a related field.
Pay
USD 223,000.00-$325,000.00 base salary (Palo Alto, CA); USD 223,000.00-$325,000.00 base salary (Seattle, WA); USD 223,000.00-$325,000.00 base salary (New York, NY)
Tools
PyTorch, TensorFlow, Kubernetes, GPU
AI focus
Build enterprise agentic and generative AI platform capabilities, specialized agents, model-serving and orchestration patterns, and evaluation methods.
Model-written summary · the original posting is authoritative