Agentic Interface Lead, Applied AI

JPMorgan Chase JPMorgan Chase · Banking · LONDON, United Kingdom · Commercial & Investment Bank

Lead the design and delivery of human-in-the-loop (HITL) interfaces for supervising and steering AI agents in financial operations, focusing on production-grade web applications with React and TypeScript, and integrating with agentic workflows.

What you'd actually do

  1. Design and build HITL interfaces for supervising, steering, and intervening in agentic workflows
  2. Prototype new interaction patterns for emerging agentic capabilities and transition successful prototypes to production
  3. Collaborate with UX as a creative peer, shaping both interaction design and implementation
  4. Set technical direction for the HITL stack, including framework selection, state management, and streaming UI integration
  5. Define contracts between agent runtimes and the supervision surface in partnership with backend and model engineers

Skills

Required

  • Expertise in building and maintaining production-grade web applications at scale
  • Advanced proficiency with modern React and TypeScript
  • Hands-on experience with CopilotKit, AG-UI, Vercel AI SDK, LangGraph UIs, or similar frameworks
  • Strong design sensibility and ability to collaborate with designers as a creative partner
  • Understanding of LLM-powered UI realities, including streaming, tool call rendering, structured outputs, latency, and context limits
  • Ability to take ideas from prototype to shipped code
  • Effective communicator across technical and non-technical teams

Nice to have

  • Experience with human-in-the-loop ML systems beyond LLMs (e.g., active learning, annotation tooling, expert review workflows)
  • Background in financial services or regulated industries with exposure to operational workflows
  • Contributions to open source in agentic UI or AI SDK domains
  • Familiarity with MCP (Model Context Protocol) and tool gateways
  • Formal training in HCI or a design-related field

What the JD emphasized

  • production-grade web applications at scale
  • modern React and TypeScript
  • CopilotKit, AG-UI, Vercel AI SDK, LangGraph UIs, or similar frameworks
  • Understanding of LLM-powered UI realities, including streaming, tool call rendering, structured outputs, latency, and context limits
  • Ability to take ideas from prototype to shipped code

Other signals

  • building interfaces that empower people to supervise and steer fleets of intelligent agents
  • own the design and delivery of interfaces that enable human supervision of AI agents at scale
  • design and build HITL interfaces for supervising, steering, and intervening in agentic workflows
  • prototype new interaction patterns for emerging agentic capabilities and transition successful prototypes to production
  • Ensure interfaces support approvals, corrections, exceptions, streaming reasoning, tool calls, and multi-agent coordination