Preference Model

Builds reinforcement-learning training environments for AI labs developing and aligning advanced models

+ Add to sales board

Highlights

  • New on EvolcAdded 2026-11-14

About Preference Model

Preference Model is a data research company that builds reinforcement-learning environments for training AI models. Its work includes environments and graders designed to resist reward hacking, along with post-training and data-curation tasks.

The company is based in San Francisco, where its listed job openings are located. Its team includes co-founders Jennifer Zhou and Ning Cao, and it is backed by South Park Commons, Scale Angel Group, Manifund, MoE Capital, and individual investors.

Product: Reinforcement-learning environments, graders, and related data research for training and post-training AI models.

Jobs at Preference Model

No open positions listed right now.

More ai & ml companies

  • Preql AI

    AI data-cleaning and semantic-modeling software for finance teams and complex enterprises
    United StatesAI & ML
    New on Evolc
  • PressW

    AI engineering and managed AI services for private equity firms and their portfolio companies
    Austin, United StatesAI & ML
    New on Evolc
  • Prime Intellect

    AI model training, inference and GPU compute platform for teams building and deploying custom models
    United StatesAI & ML10 open jobs
    Remote-friendly
  • Primordial Labs

    Natural-language control software for defense operators coordinating uncrewed systems, sensors and mission software
    United StatesAI & ML
    New on Evolc
  • Probook

    AI dispatch and customer-scheduling software for home services operators
    United StatesAI & ML
    New on Evolc
  • Procurement Sciences

    AI software for government contractors to find opportunities and manage capture, proposals and business development
    Washington, United StatesAI & ML
    New on Evolc

All ai & ml companies →

Website technologies

Tools and platforms detected on this company’s website, grouped by purpose. Dates record when Evolc observed them.

Frontend frameworks

  • Next.jsHigh confidence
    First detected
    Last detected
    Last checked
    Detection evidence
  • ReactMedium confidence
    First detected
    Last detected
    Last checked
    Detection evidence

Hosting

  • AWSHigh confidence
    First detected
    Last detected
    Last checked
    Detection evidence

Recruiting

  • AshbyMedium confidence
    First detected
    Last detected
    Last checked
    Detection evidence

Website detection does not establish a paid subscription or company-wide use. Tools loaded after consent or sign-in may be missed. “No longer observed” does not prove the company stopped using a tool.