Research Engineer / Research Scientist, RL Frontiers
Core
Develop next-generation RL algorithms and model architectures, scaling them from small experiments to frontier-scale runs for Anthropic's AI systems.
Role type
Senior IC research engineer (RL scaling)
Builds
Next-gen RL algorithms, model architectures, and experimental infrastructure for large-scale AI training
Domain
AI / Reinforcement Learning / Large-scale Systems
Deliverable
production ML models
Required skills
Transformer architecture expertise, distributed model training, experimental design at scale, quantitative compute modeling, Python/JAX/PyTorch, systems debugging
Preferred skills
RL research experience, scaling laws modeling, transformer architecture modification, large fleet debugging, low-precision numerics, C++/Rust
Technologies
JAX, PyTorch, Python, C++, Rust, GPU/TPU clusters
Responsibilities
Study RL training scaling with model size and compute; develop frontier-scale algorithms; diagnose scale-induced instabilities; build reproducible experimental infrastructure; own end-to-end performance of largest RL runs; model throughput and cost for architecture changes.
Seniority
Senior, hands-on IC
