Principal Machine Learning Engineer, Agent Harness - Meta Factory

Adobe Adobe · Enterprise · San Jose, CA

Principal Machine Learning Engineer to define the technical architecture of the Agent Harness, a platform for autonomous agents that understand builder intent, decompose work, use tools, evaluate outcomes, and improve over time. This role involves setting technical direction, solving complex design challenges, and establishing engineering standards for AI-native software development at Adobe scale. Focus on agent execution, tool use, context, state, and lifecycle management, while balancing performance, cost, and reliability. Requires expertise in distributed systems, runtime build, platform architecture, agentic systems, and LLM infrastructure, with hands-on experience in fine-tuning, optimizing, and operating models or runtimes at scale.

What you'd actually do

  1. Work as part of a team of architects shaping the technical direction of the Agent Harness, contributing across agent execution, tool use, context, state, and lifecycle management.
  2. Define harness interfaces and standards while balancing performance, cost, and reliability.
  3. Track emerging trends in agent runtimes and tool-use standards and lead their evaluation and adoption at Adobe.
  4. Define AI-first architecture and engineering standards for agentic systems across Adobe.
  5. Lead cross-team technical initiatives and provide guidance on complex architectural challenges.

Skills

Required

  • 15+ years of software engineering experience
  • Deep expertise in distributed systems, runtime build, and platform architecture
  • Recognized expertise in agentic systems and LLM infrastructure
  • hands-on experience fine-tuning, optimizing, and operating models or runtimes at scale
  • Experience driving AI-first architecture, engineering standards, technology choices, and product direction across organizations
  • Strong communication and influence skills

Nice to have

  • leading large multi-functional initiatives
  • align senior leaders and teams around a shared technical vision

What the JD emphasized

  • define the technical architecture
  • senior technical leaders in agentic systems
  • set the long-term technical direction
  • solve the hardest design challenges
  • establish engineering standards for AI-native software development at Adobe scale
  • agent execution
  • tool use
  • context
  • state
  • lifecycle management
  • agent runtimes
  • tool-use standards
  • AI-first architecture
  • engineering standards for agentic systems
  • complex architectural challenges
  • Deep expertise in distributed systems, runtime build, and platform architecture
  • leading large multi-functional initiatives
  • Recognized expertise in agentic systems and LLM infrastructure
  • hands-on experience fine-tuning, optimizing, and operating models or runtimes at scale
  • Experience driving AI-first architecture, engineering standards, technology choices, and product direction across organizations
  • Strong communication and influence skills
  • align senior leaders and teams around a shared technical vision

Other signals

  • agent execution
  • tool use
  • context
  • state
  • lifecycle management
  • LLM infrastructure
  • agentic systems