Principal Machine Learning Engineer, Agentic AI

Zillow Zillow · Consumer · United States · Remote

The Principal Machine Learning Engineer on the Agentic AI team at Zillow will develop advanced multimodal foundational technologies, focusing on prototyping, evaluating, and deploying models into production. This role involves building multi-agent systems with deep reasoning, perception, and language understanding capabilities, leveraging frameworks like AgentSDK and LangChain/LangGraph. The engineer will also work with GenAI models, real-time voice APIs, and ensure high availability and low latency for AI services.

What you'd actually do

  1. Leverage frameworks like AgentSDK, and LangChain/LangGraph to design, prototype, and develop multi-agent systems that are capable of highly autonomous and context-aware interactions
  2. Leverage advanced GenAI models including reasoning models, real-time voice API, etc, to build agentic prototypes and later on convert them into product-level agentic skills and deploy to users.
  3. Have the mentality of “build, learn, and pivot”, and hold the high bar of shipping into production.
  4. Mentor and guide engineers in using the right technologies for agentic AI solutions and foster a culture of innovation and responsible AI usage
  5. Distill complex research findings and system designs into actionable insights for diverse audiences—including executives

Skills

Required

  • Master's degree or above, equivalent experience in Computer Science, Electrical Engineering, or a related field with emphasis on foundational LLM, agentic AI, reinforcement learning, AI planning, or natural language processing
  • 7+ years of hands-on work building large-scale, high-impact solutions—ideally in the most recent two years building agent-based systems, multi-agent collaboration, or similar paradigms
  • Experience developing dialogue systems capable of long conversations, multi-step reasoning, context-rich decision-making
  • Experience deploying and scaling AI services capable of handling hundreds of millions of daily interactions with high availability, low latency, and robust fault tolerance

Nice to have

  • A track record of publishing high-impact research in top AI/ML venues is a big plus.

What the JD emphasized

  • shipping into production
  • deploying and scaling AI services

Other signals

  • develop advanced multimodal foundational technologies
  • prototyping ideas, evaluating, and deploying models
  • develop perception and language understanding, deep reasoning, and reinforcement learning
  • design, prototype, and develop multi-agent systems
  • build agentic prototypes and later on convert them into product-level agentic skills and deploy to users
  • shipping into production
  • applied science projects
  • building agent-based systems, multi-agent collaboration
  • deploying and scaling AI services