Principal Product Manager, Agentic AI Platform

Zillow

2d•$185,800 - $312,400•Remote

About The Position

The Principal Product Manager for the Agentic Platform is the technical product leader responsible for how Zillow teams build and ship AI agents on a shared platform. As part of the Product AI team, this individual works with product managers, platform engineers, scientists, and Core Tech, connecting the core agent platform and data platforms with experience teams across consumer shopping, seller, rentals, agent and pro tools, and mortgage. This is an individual contributor role reporting to the leader of Product AI, with impact achieved through influence, technical credibility, and shaping a platform that enables teams to create experiences they could not build alone. The role shapes the platform's technical product direction across agent and tool registration, discovery and governance; developer experience and team onboarding; user and domain context and memory, including privacy and user controls; and the shared agent runtime, orchestration, and evaluation standards. The Principal Product Manager works closely with product, engineering, and science partners to make these capabilities fit together as a coherent platform and customer experience. This position is categorized as remote. Remote employees do not have a permanent corporate office workplace and work from a physical location of their choice that must be identified to the Company. U.S. employees may live in any of the 50 states, subject to limited exceptions. The role is eligible for equity awards in addition to base salary. The stated annual base salary range is $195,600–$312,400 in certain listed locations and $185,800–$296,800 in another listed group of locations; actual compensation varies based on experience, performance, and location.

Requirements

Bachelor’s degree in Computer Science, Computer Engineering, Data Science, or a related technical field.
10 or more years of product management experience, including platform or infrastructure products serving other product and engineering teams as customers.
Hands-on depth with LLM and agentic systems, including orchestration, context engineering, memory and RAG, and evaluation design.
Track record of driving adoption through influence across organizations with different incentives and roadmaps.
Ability to write concise, compelling, decision-forcing documents for SVP-level audiences and lead the execution that follows.
Comfort with ambiguity and contested ownership, and ability to resolve scope disputes by redrawing boundaries rather than escalating.
Ability to bring engineering, science, and domain teams together and turn differing viewpoints into better decisions.

Nice To Haves

Data mesh or federated data platform experience.
Experience with experimentation and A/B testing on AI features.
Consumer marketplace or two-sided (consumer plus professional) product experience.

Responsibilities

Shape the platform roadmap with product, engineering, and science partners, translating product and customer needs into clear technical requirements.
Run a proving-team operating model by onboarding two to three flagship experience teams, defining graduation criteria, and scaling to subsequent teams.
Make open architecture decisions and ensure each is closed with a named decision-maker and a date.
Turn cross-organizational dependencies into explicit commitments, including named owners on both sides, capacity range, re-plan date, and exit criteria.
Bridge traditional engineering with agentic science and data needs, helping teams resolve source-of-truth debates, one-way-door schema decisions, facts-versus-inferences questions, and the hill-climbing process.
Define and instrument platform success, including reuse, time-to-first-agent, cost per team, and coherence of the experience across surfaces.
Maintain and openly share a clear list of what the platform is not building.
Shape how agents, tools, and context sources are registered, discovered, and governed, including contract adherence at the edge.
Develop paved paths, reference integrations, and onboarding that reduce the time it takes a new team to build its first agent.
Shape how user and domain context is captured, curated, and delivered into agent context windows, with privacy and user controls built in.
Partner with product managers and engineers to ensure the shared agent runtime and quality standards work together across the customer experience.