6 months ago
Responsibilities
- Conduct a full-time empirical research or engineering project aligned with Anthropic’s research priorities
- Build CPU simulators for accelerator workloads
- Add backends for different accelerators to open-source projects
- Build on-demand infrastructure for infrastructure-heavy research projects
- Develop complex synthetic data or environment pipelines
- Analyze and debug model training processes
- Collaborate with Anthropic researchers and engineering and research disciplines
- Produce a public output such as a paper submission
Requirements
- Strong technical background in computer science, mathematics, physics, or a related discipline
- Fluency in Python programming
- Availability to work full-time for the four-month program
- Strong software engineering skills and experience building complex ML systems are particularly valuable
- Experience with large-scale distributed systems and high-performance computing is particularly valuable
- Experience training, fine-tuning, or evaluating large language models is particularly valuable
- Ability to analyze and debug model training processes
- Work authorization in the US, UK, or Canada and the ability to be located in one of those countries during the program
- Relevant research or engineering experience is helpful but not required
Benefits
- Four-month full-time program with possible extension
- Weekly stipend plus country-specific benefits
- Direct mentorship from Anthropic researchers
- Access to shared workspaces in Berkeley or London
- Remote work available in the US, UK, or Canada
- Funding for compute and other research expenses
- Flexible workspace and collaboration with the broader AI safety and security research community
- Work authorization is required in the US, UK, or Canada; visa sponsorship is unavailable
- Next cohort starts July 20, 2026, with applications due April 26, 2026
Tech Stack
Categories
About Anthropic
Anthropic builds large language models and the Claude AI assistant for developers and enterprises, offered via API access and enterprise plans. Founded in 2021 and headquartered in San Francisco, it distributes Claude through its own platform and via partners such as Amazon Bedrock and Google Cloud’s Vertex AI. Its work emphasizes model reliability, interpretability, and practical tooling for tasks like coding assistance, analysis, and customer support automation.