Lead User Modeling Scientist (RL & Causal Inference)

ZINC Zillow, Inc.; United States; Remote; USD 186,000 - 297,000; Full time

Benefits offered by this job

Equity awards

Job summary

Zillow Group is seeking an Applied Scientist to advance user modeling within the Agentic Foundations team. You will own end-to-end training pipelines and build reward models, leveraging state-of-the-art sequence modeling and reinforcement learning to impact real home-shopping experiences at scale.

The role emphasizes collaboration across product, engineering, science, and data teams, with remote flexibility and equity opportunities.

Qualifications

• PhD in a quantitative field with depth in RL or causal inference.
• Hands-on experience modeling user behavior at scale with sequence models.
• Proficiency in Python and ML frameworks to run experiments and review code.

Responsibilities

• Own LLM post-training pipelines (SFT, DPO, RFT/GRPO) on GPU infrastructure.
• Build and train fast, multi-category reward models emitting scalar signals.
• Design on-policy and online assessment to score live agent trajectories.
• Translate offline evaluation rubrics into generalizable reward functions for production traces.
• Independently drive complex work end-to-end across partner teams.
• Provide technical leadership and mentorship translating research into production impact.

Skills

Python

PyTorch

TensorFlow

Reinforcement learning

Causal inference

User modeling

Education

PhD in Computer Science/Statistics/Economics/ML

Job description

Zillow Group is seeking an Applied Scientist to advance user modeling within the Agentic Foundations team. You will own end-to-end training pipelines and build reward models, leveraging state-of-the-art sequence modeling and reinforcement learning to impact real home-shopping experiences at scale.

The role emphasizes collaboration across product, engineering, science, and data teams, with remote flexibility and equity opportunities.