Research Engineer, Machine Learning (Reinforcement Learning)

• Join Anthropic’s Reinforcement Learning teams as a Research Engineer in London. • Help advance the capabilities and safety of large language models, including Claude. • Blend research and engineering, implementing novel approaches and shaping research direction. • Build agentic models that use tools for computer use and autonomous software generation. • Improve model reasoning in areas such as mathematics. • Earn £260,000–£630,000, working from the London office at least 25% of the time.

• Write Python proficiently, including async and concurrent programming with frameworks such as Trio. • Bring experience with machine learning frameworks such as PyTorch, TensorFlow or JAX. • Show industry experience in machine learning research. • Balance research exploration with engineering implementation. • Enjoy pair programming. • Care about code quality, testing and performance. • Bring strong systems design and communication skills. • Feel committed to developing safe and beneficial AI. • Hold a bachelor’s degree or an equivalent mix of education, training and experience.

• Know LLM architectures and training methodologies. • Bring experience with reinforcement learning techniques and environments. • Show experience with virtualisation and sandboxed code execution. • Bring experience with Kubernetes. • Show experience with distributed systems or high-performance computing. • Write Rust or C++.

• Architect and optimise core RL infrastructure, from training abstractions to distributed experiments across GPU clusters. • Design, build and test novel training environments, evaluations and methodologies for RL agents. • Improve performance across the stack through profiling, optimisation and benchmarking. • Implement efficient caching and debug distributed systems to speed up training and evaluation. • Build automated testing frameworks, clean APIs and scalable research infrastructure. • Develop prototypes for internal use, productivity and evaluation. • Collaborate with alignment, frontier red team and production training teams.

• Earn £260,000–£630,000 a year. • Receive visa sponsorship, supported by a retained immigration lawyer. • Enjoy generous vacation and parental leave. • Work flexible hours with a hybrid setup of at least 25% in the office. • Get optional equity donation matching. • Work on large-scale research that shapes frontier AI.

Anthropic is a public benefit corporation headquartered in San Francisco, with a mission to create reliable, interpretable and steerable AI systems. It works as a single cohesive team on a few large-scale research efforts, treating AI research as an empirical science. Its team includes researchers, engineers, policy experts and business leaders building AI that is safe and beneficial for society.