AI Research Scientist, Recursive Self Improvement, AI Safety and Reinforcement Learning
Core
Research recursive self-improvement (RSI) systems, AI safety, and reinforcement learning to ensure safe, auditable training loops for AMD's AI-for-HW and generative-AI programs.
Role type
Senior AI Research Scientist (AI Safety & RL)
Builds
Safe, bounded self-improving training pipelines and evaluation frameworks
Domain
Artificial Intelligence, AI Safety, Reinforcement Learning
Deliverable
production ML models
Required skills
Machine learning, AI safety, reinforcement learning, empirical safety evaluation, scalable oversight, software engineering for experimental harnesses
Preferred skills
Publications in iterative training/self-training, stress-testing generative model pipelines, building reproducible RSI microcosms
Technologies
None explicitly listed
Responsibilities
Research self-improving training loops (model-generated supervision, iterative distillation, self-critique, automated curriculum updates); Develop evaluations for capability drift, Goodhart effects, and distributional shift; Partner with RL scientists on policy optimization and preference learning; Define red-team protocols and monitoring for RSI pilots; Publish technical reports and align with responsible deployment standards
Seniority
Senior, hands-on IC
