Senior Staff Full Stack Engineer - Agent Platform, Firefly Boards

Adobe Adobe · Enterprise · San Jose, CA

Senior Staff Engineer to lead the Agent Platform for Firefly Boards, an AI-powered creative app. This role owns the production-grade engineering systems for the server-side agent runtime, integration with the central agent platform, and the application stack enabling real-time agent actions on an infinite canvas. Responsibilities include state management, checkpointing, error recovery, observability for long-running operations, defining the browser/server execution boundary for agent tools, and integrating with Adobe's central agent platform. The role also involves refactoring APIs for agent consumption and ensuring production-grade systems concerns like observability and latency budgets. This is a senior IC role focused on engineering direction and cross-organizational partnerships.

What you'd actually do

  1. Server-side agent runtime. Stand up and own the service that runs Boards' agent workflows. State management, checkpointing, interrupt-and-resume across long-running operations, error recovery, and observability. This is the spine of the system and the foundation everything else builds on.
  2. Browser/server execution boundary. Define and own how server-side workflows dispatch to browser-only tools (canvas operations, viewport state, selection) and resume cleanly. Design the protocol, the runtime contract, and the in-client tool registration and invocation surface. Make it reliable under multiplayer conditions where humans and agents are editing the same document.
  3. Agent platform integration. Own the integration between Boards' agent runtime and Adobe's central agent platform. Co-shape platform contracts where Boards needs services that don't yet exist in the platform-supported shape. This is a cross-org partnership, not a one-off integration.
  4. Boards backend services for agentic consumption. Refactoring existing Boards APIs and adding new ones to be agent-callable. API design under real constraints: idempotency, transactional boundaries, undo, multiplayer-aware mutations.
  5. Production-grade systems concerns. End-to-end observability and tracing for agent runs. Latency budgets across runtime, tools, and model calls. Deployment, rollout, rollback. The work that turns a working prototype into something that's operable at scale.

Skills

Required

  • 10+ years of product engineering experience
  • significant ownership of production systems at scale
  • Demonstrated full stack depth: backend services and modern web client work
  • JavaScript/TypeScript
  • declarative UI frameworks (React, Lit, Vue)
  • strong server-side language
  • Distributed systems depth
  • state management
  • checkpointing
  • async work
  • long-running operations
  • recovery from partial failure
  • designing and operating systems with these properties
  • Experience integrating with a central platform team as a tenant
  • Experience designing service APIs and microservices
  • REST or equivalent contracts
  • Experience driving sophisticated software architecture, design, and development across multiple teams or organizations
  • Track record of writing clean, testable code
  • creating reusable components
  • Strong written and verbal communication
  • Comfortable shaping technical decisions in cross-org forums

Nice to have

  • Experience building agentic or retrieval-augmented (RAG) systems with workflow engines such as LangGraph, or comparable runtime orchestration for tool-using LLM applications
  • Experience with real-time collaborative or multiplayer applications, especially the browser/server boundary in a collaborative editor
  • Familiarity with MCP (Model Context Protocol) and agent tool design patterns
  • Experience with production LLM systems: evaluation infrastructure, observability, latency optimization, prompt and context engineering at the system level
  • Experience with collaborative document models or CRDT-based systems
  • Experience building image, video, or multimedia applications
  • Experience with development tools such as Jenkins, GitHub, lint, etc.
  • Experience with improving performance and stability of large web applications
  • Experience with Accessibility (A11y) and Localization
  • Experience working in an Agile development environment

What the JD emphasized

  • production-grade
  • server-side agent runtime
  • agent platform integration
  • production-grade systems concerns
  • end-to-end observability and tracing for agent runs
  • latency budgets across runtime, tools, and model calls
  • long-running operations
  • recovery from partial failure
  • real-time collaborative or multiplayer applications

Other signals

  • agentic AI
  • server-side agent runtime
  • integration with Adobe's central agent platform
  • application stack that lets the agent act on the canvas
  • production-grade systems concerns
  • end-to-end observability and tracing for agent runs
  • latency budgets across runtime, tools, and model calls