Summary

As a member of technical staff on the safety for agents team, you will work on data generation, post-training algorithms, and evaluation methods to improve the safety of large language models that can access external resources and take actions. You will collaborate with cross-functional machine learning teams, data annotation teams, and product and policy teams, with a remote-friendly setup requiring ~50% overlap with uk/eu timezones.

Responsibilities

• Generate safety-focused training data
• Develop post-training algorithms for llms
• Design evaluation methods for safety
• Analyze datasets for quality and biases
• Conduct data collection tasks with annotators
• Train llms on distributed training infrastructure
• Assess generalizability and robustness