Research Scientist, STEM
Core
Develop new ways to evaluate, train, and improve frontier AI systems through research on benchmarks, synthetic data, hallucination, and agentic science.
Role type
Research Scientist (Frontier AI)
Builds
Novel benchmarks, synthetic STEM training data, evaluation methodologies, and agentic science workflows.
Domain
Artificial Intelligence / Frontier AI Research
Deliverable
research
Required skills
Machine learning, experimental design, quantitative reasoning, Python, technical writing, rapid literature synthesis
Preferred skills
PhD in ML/CS/Math/Physics/Chemistry/Biology/Engineering/Statistics, publication record, experience with LLMs
Technologies
Python, LLMs, Reinforcement Learning environments
Responsibilities
Identify gaps in benchmark literature and design novel STEM benchmarks; Develop methods for generating high-quality synthetic STEM training data; Study hallucination, uncertainty, and verification in technical domains; Research AI systems capable of extended scientific and technical work (agentic science).
