We launched recruiters.rip, let your AI apply for you. First month free with code onemonthfree

Research Engineer, Beneficial Deployments

AI safety and research company building reliable, interpretable, and steerable AI systems, including the Claude product family and developer platform.

Series F+Recently funded59 current maintainers55 active leads23 new active leads26 lead step-downs8 early lead departures
San Francisco, United States
About Anthropic

Anthropic PBC develops frontier AI systems and deploys them through Claude products and the Claude Platform, with a stated focus on safety, interpretability, and steerability.

View jobs by Anthropic

Skills

Candidate Availability

About the Role

You will identify where Claude falls short, build the RL environments and reward signals used to train it, run generalization experiments, and help shape the Beneficial Deployments roadmap based on where AI capabilities are heading.

Requirements

  • Experience fine-tuning large language models
  • Experience with reinforcement learning, reward design, or training data curation for LLMs
  • Track record of self-directed delivery on ambiguous high-stakes work
  • Ability to turn technical understanding into strategic recommendations for non-technical leaders
  • Comfort reading datasets to spot issues
  • Preferred: experience training production ML systems, designing evals, healthcare or global health expertise, multilingual models

Responsibilities

  • Identify capability gaps in models and turn them into a research agenda
  • Own end-to-end creation of RL environments from task sourcing and reward design through training
  • Collaborate with domain experts to design data pipelines
  • Develop QA frameworks to catch reward hacking and ensure environment quality
  • Run generalization experiments
  • Partner with RL research and product teams

Benefits

  • Competitive compensation and benefits
  • Optional equity donation matching
  • Generous vacation and parental leave
  • Flexible working hours
  • Visa sponsorship