Research Engineer, Environment Scaling

Anthropic is a public benefit corporation focused on creating reliable and beneficial AI systems. The Research Engineer for Environment Scaling will work on improving the intelligence of public models by building training environments, managing vendor relationships, and executing ML research to enhance model capabilities.

Responsibilities:

• Improve and execute our fine-tuning strategies for adapting Claude to new domains and tasks
• Manage technical relationships with external data vendors, including evaluation of data quality and reward design
• Collaborate with domain experts to design data pipelines and evaluations
• Explore novel ways of creating RL environments for high value tasks
• Develop and improve QA frameworks to catch reward hacking and ensure environment quality
• Partner with other RL research teams and product teams to translate capability goals into training environments and evals

Requirements:

• Bachelor's degree or an equivalent combination of education, training, and/or experience
• A field relevant to the role as demonstrated through coursework, training, or professional experience
• Experience with fine-tuning large language models for specific domains or real-world use cases and/or domain expertise in an area where we would like to make our models more useful
• Experience with reinforcement learning, reward design, or training data curation for LLMs
• Comfortable managing technical vendor relationships and iterating quickly on feedback
• Value in reading through datasets to understand them and spot issues
• Strong project management and interpersonal skills
• Passionate about making AI more useful and accessible across different industries
• Excited about a role that includes a combination of ML research, data operations, and project management
• Experience training production ML systems
• Familiar with distributed systems and cloud infrastructure
• Domain expertise in an area where we would like to make our models more useful
• Experience working with external vendors or technical partners