Research Engineer, Performance RL (Reinforcement Learning)

at Anthropic

📍 San Francisco, United States

USD 350,000-850,000 per year

MIDDLE

✅ Hybrid

✅ Visa Sponsorship

Tech Stack
Tag name is followed by "@" symbol and proficiency level value. About proficiency levels:

1-2 — basic awareness. Minimal hands-on experience, and a rudimentary understanding of the technology's purpose;

3-6 — daily use. Comfortable and regular usage, capable of handling common tasks and challenges related to the technology;

7-9 — you are an expert, you can teach others, you know all the pitfalls and tricks;

10 — exceptional knowledge, comprehensive understanding, and adeptness in all aspects of the technology, including advanced problem-solving. Think twice before claiming or demanding such level.

AI @ 3 CUDA @ 3 Distributed Systems @ 3 JAX @ 3 LLM Machine Learning @ 3 PyTorch @ 3 Reinforcement Learning @ 3

Details

Anthropic is hiring a Research Engineer for the Code RL team within its Reinforcement Learning organization. The role focuses on advancing models' ability to safely write correct, fast code for accelerators and delivering research innovations into large-scale training runs.

Responsibilities

Invent, design, and implement reinforcement learning environments and evaluations.
Conduct experiments and help shape the research roadmap.
Deliver research outcomes into training runs.
Collaborate with researchers, engineers, and performance engineering specialists across and outside Anthropic.
Develop systems and methodologies supporting reinforcement learning research for large language models, including code generation and model reasoning.

Requirements

Expertise with accelerators such as CUDA, ROCm, Triton, or Pallas.
Experience with machine learning framework programming using JAX or PyTorch.
Experience working across the stack, including kernels, model code, and distributed systems.
Ability to balance research exploration with engineering implementation.
Passion for AI's potential and commitment to developing safe and beneficial AI systems.
Bachelor's degree or an equivalent combination of education, training, and/or experience.
A relevant field of study demonstrated through coursework, training, or professional experience.

Preferred Qualifications

Experience with reinforcement learning.
Experience porting machine learning workloads between different types of accelerators.
Familiarity with large language model training methodologies.

Compensation

Annual salary: $350,000–$850,000 USD

Benefits

Anthropic offers competitive compensation and benefits, optional equity donation matching, generous vacation and parental leave, flexible working hours, and an office space for collaboration.

Logistics

Staff are currently expected to work from one of Anthropic's offices at least 25% of the time, although some roles may require more office time.
Anthropic sponsors visas when possible and makes reasonable efforts to support visa applications with the help of an immigration lawyer.
Anthropic is headquartered in San Francisco and is a public benefit corporation.