About the job
Zillow is looking for a Senior Applied Scientist on the Applied Reasoning team within the Agentic AI organization to research and develop advanced LLM-powered reasoning systems specialized in the real estate domain. In this role, you'll take a hands-on approach to prototyping, evaluating, and deploying agents capable of deep, multi-step reasoning, accurate use of tools/skills, and contextual decision-making tailored to real estate use cases.
Responsibilities
Design, prototype, and build advanced agentic systems capable of highly autonomous, context-aware, and adaptive interactions across diverse real estate use cases
Apply test-time scaling and post-training techniques to develop agents that can reason, collaborate, compete, or negotiate in dynamic, goal-driven environments to fulfill user needs.
Define and refine evaluation and experimentation processes for LLM-driven applications.
Stay at the forefront of agentic AI research and innovation, bringing emerging techniques into practical application to shape product direction.
Contribute to the broader scientific community through publications, conference presentations, and internal knowledge sharing
Mentor and guide junior scientists and engineers, promoting best practices in applied research, scalable agentic architecture, and responsible AI development.
Qualifications
Minimum
A Ph.D. or equivalent experience in Computer Science or a related field, with a focus on LLM, Agentic System, or Machine Learning
2+ years of experience in building large-scale, high-impact ML solutions, particularly in areas such as NLP, agent-based systems, or multi-agent collaboration or similar paradigms
Experience in developing and working with LLM reasoning models or AI agents capable of multi-step reasoning and context-rich decision-making
Strong background in rigorous experimental design and evaluation, including benchmark creation, ablation studies, qualitative and quantitative analysis, and principled measurement of reasoning quality, generalization, and safety
Preferred
Track record of publishing high-impact research in top AI/ML venues