Distinguished Scientist

About the team

The Agentic AI team at Zillow is at the forefront of transforming the real estate industry by helping millions of people use advanced AI technologies across every step of their home-shopping and financing journey. We are building proactive, always-on assistants that integrate deep user understanding, market intelligence, and sophisticated reasoning to guide people from initial intent through closing and beyond. By embedding domain knowledge and specialized expertise into these assistants, we empower industry professionals and consumers to work more efficiently while preserving Zillow’s competitive edge and trusted brand.

Our team brings together applied scientists, machine learning engineers, and software engineers who own the full lifecycle of AI experiences—from data collection and foundation model training to agentic skill design, multi-agent orchestration, evaluation, and large-scale deployment. We operate in a highly collaborative environment with AI platform, finance, and product teams, and we rely on Zillow’s unique, privacy-preserving real-estate and transaction datasets to build capabilities that few other companies can match.

In this role, you will define and drive the technical vision for long-horizon, supervised agentic systems  will design agent architectures, memory and supervision frameworks, and learning strategies that allow our agents to build and maintain context-rich relationships over weeks and months, not just single sessions, while remaining trustworthy, compliant, and measurably helpful. You will leverage Zillow’s uniquely rich domain datasets—such as interaction logs, behavioral signals, and structured real-estate and transaction data—to train and adapt agents that make high-quality, human-caliber recommendations at scale.

As a Distinguished Scientist, you will:

• Own the end-to-end research and technical strategy for long-horizon agentic experiences across shopping, financing, and professional workflows, in close partnership with the Agentic AI, data platform, and product teams.
• Design and advance LLM post-training and evaluation methods (e.g., SFT, preference learning, RLHF/RLAIF, long-context modeling) tailored to supervised, high-stakes journeys in a complex, regulated domain.
• Architect systems that combine persistent memory, tool use, and multi-agent collaboration to deliver consistent, context-rich guidance over long timelines.
• Translate Zillow’s heterogeneous data (text, voice, behavioral, and structured real-estate/transaction data) into agent-ready knowledge and signals, in partnership with data and platform teams.
• Collaborate with product and design to define success metrics, evaluation frameworks, and experiment plans for agentic experiences, including human-in-the-loop supervision and safety reviews.
• Operate as a senior IC with the option to lead a small pod (up to ~5 scientists/engineers) focused on long-horizon agentic systems, mentoring principal-level talent and setting a high technical bar.
• Represent Zillow’s work externally through publications, talks, open-source contributions, and thoughtful engagement with the research community, helping position Zillow as a destination for top agentic AI talent.

This is a highly visible role with broad impact across AI & Experiences. You will partner closely with senior leaders to align foundational research, platform investments, and product roadmaps around a coherent vision for Zillow as an AI-native company.

The responsibilities are covered in the bullet list above.

You have:

• A PhD in Computer Science, Electrical Engineering, or a related field—or equivalent experience—with emphasis in areas such as foundational LLMs, agentic AI, reinforcement learning, AI planning, or natural language processing.
• 10+ years of hands‑on experience building and deploying large-scale AI systems, including at least several years focused on agent-based systems, multi-agent collaboration, or long-horizon conversational assistants.
• Deep, current expertise in generative and agentic AI, including multimodal foundation models, transformers, advanced reasoning models, and post‑training…