Research Scientist
Core
Conduct original technical research in AI safety to build foundations for safe, trustworthy AI deployment, designing experiments to study and steer model behavior.
Role type
Research Scientist (AI Safety)
Builds
Robust, reproducible research code and technical tooling for AI safety
Domain
Artificial Intelligence / AI Safety
Deliverable
production ML models
Required skills
Machine learning fundamentals, transformer architectures (LLMs), white-box and black-box interpretability, Python programming, deep learning frameworks (PyTorch), Git version control, hypothesis formulation, experimental design
Preferred skills
Experience in AI safety areas (robustness, evaluations, uncertainty calibration), ability to translate technical insights for external partners
Technologies
Python, PyTorch, Git
Responsibilities
Design and run experiments to study and steer model behavior; Develop modular and reproducible research code; Build and maintain research infrastructure; Document findings via academic papers and technical reports; Collaborate with senior scientists to refine hypotheses; Translate complex technical insights for external audiences
Seniority
Mid-level IC