Taste Labs
AI Engineer, RL & Evals
$175K–$300K
San Francisco · On-site · Full-time · Posted Aug 20, 2026
About the job
Build evaluation methods, RL environments, and infrastructure that help frontier AI models develop and demonstrate taste in design.
What you'll do
- Research grading methods and rubrics, and design tasks that assess taste and design capabilities.
- Build agent harnesses, context layers, and scalable RL infrastructure; support internal training pipelines.
- Collaborate with frontier labs to craft environments that improve their models.
What they ask for
- Experience building evaluations, RL environments, ML, or post-training.
- Strong backend experience.
- Ability to tackle ambiguous, difficult, creative problems and make subjective domains verifiable.
- Team-oriented, adaptable, fast-moving, and willing to take ownership.
Nice to have
- Open-source contributions or personal projects demonstrating curiosity-driven building.
- Background at a creative, index-building/crawling, or data company.
Pay
$175K – $275K USD base salary; Up to $300K USD base salary for exceptional candidates
AI focus
Develop evaluations, post-training data, RL environments, and agent tools to improve AI models’ judgment and taste, especially in design.
Model-written summary · the original posting is authoritative