Sr Staff Agentic Systems Engineer

Uber Uber · Consumer · Seattle, WA +2 · Engineering

Uber is seeking a Sr Staff Agentic Systems Engineer to build the infrastructure for Agentic Dev Environments (ADEs), where AI agents are first-class participants in the software development lifecycle. The role involves designing and shipping the skill pack framework, building the multi-agent runtime with persistent background agents and swarm orchestration, developing AI-powered code intelligence, and owning the MCP platform layer for unified context infrastructure. The engineer will set technical direction for agentic infrastructure, define architecture and trust models, and mentor engineers in agentic fluency, working at Uber's scale across multiple monorepos and microservices.

What you'd actually do

  1. Design and ship the skill pack framework. How skills are authored, versioned, distributed, and loaded across Go, Java, Python, iOS, Android, and Web monorepos. Skills are the differentiator: portable expertise that makes any agent instantly productive.
  2. Build the multi-agent runtime inside Uber's DevPod cloud environments. Persistent background agents, swarm orchestration, cross-agent context passing. Agents working 24/7 in parallel, not one-shot interactions.
  3. Develop AI-powered code intelligence. Semantic code search, dependency-aware context, codebase understanding that gives agents deep structural knowledge (not just text search).
  4. Own and evolve the MCP platform layer. Consolidating code-mcp, docs-mcp, and developer lifecycle MCPs into a unified, low-latency context infrastructure that agents actually depend on.
  5. Partner with Code Infra, DevPod, Mobile Platform, Web Platform, and Backend Platform teams to ship the foundational systems that make agentic engineering work at Uber's scale. ~5K microservices, 6 monorepos, 5,000+ engineers.

Skills

Required

  • Experience building and scaling developer tooling or platform infrastructure (e.g., SDKs, CLIs, APIs, runtime systems) in large-scale engineering environments
  • Strong proficiency in at least one major programming language (Go, Java, Swift, Kotlin, Python, or TypeScript) with demonstrated experience integrating AI/ML models into developer workflows
  • System design experience for high-reliability, observable, developer-facing services and agentic systems

Nice to have

  • Experience building and scaling developer tooling or platform infrastructure (e.g., SDKs, CLIs, APIs, runtime systems) in large-scale engineering environments
  • Strong proficiency in at least one major programming language (Go, Java, Swift, Kotlin, Python, or TypeScript) with demonstrated experience integrating AI/ML models into developer workflows
  • System design experience for high-reliability, observable, developer-facing services and agentic systems

What the JD emphasized

  • agent orchestration system for autonomous code generation
  • multi-agent runtime
  • swarm orchestration
  • cross-agent context passing
  • AI-powered code intelligence
  • dependency-aware context
  • unified, low-latency context infrastructure
  • agentic engineering work at Uber's scale
  • agent-generated code at org scale

Other signals

  • AI agents as first-class participants in the software development lifecycle
  • building the infrastructure that makes this shift real
  • agent runtime
  • agent orchestration system for autonomous code generation
  • platform layer giving agents structured access to Uber's tools, data, and developer lifecycle
  • Skill Marketplace
  • foundational pieces of Uber's agentic infrastructure
  • ship the skill pack framework
  • multi-agent runtime
  • persistent background agents
  • swarm orchestration
  • cross-agent context passing
  • AI-powered code intelligence
  • semantic code search
  • dependency-aware context
  • codebase understanding
  • unified, low-latency context infrastructure that agents actually depend on
  • ship the foundational systems that make agentic engineering work at Uber's scale
  • Set technical direction for agentic infrastructure
  • architecture, verification frameworks, and trust models for agent-generated code
  • Mentor engineers in agentic fluency
  • build with, and ship through, AI agents