Shape safer Large Language Models through statistical evaluation and machine learning, working remotely or from Cohere offices with UK/EU overlap.
As a Senior Member of Technical Staff, you will develop safety methods for agents accessing external resources and taking actions. Your work combines ethical AI, experimental design, data generation, and software engineering. You will collaborate with machine learning, annotation, product, and policy teams.
Key Responsibilities: Generate data for safety-focused post-training algorithms
Design evaluations for model performance and robustness
Analyze dataset quality, bias, and training suitability
Train LLMs on distributed infrastructures
Present findings across cross-functional teams
Requirements: Strong statistical skills evaluating scientific experiments
Extremely robust software engineering skills
Experience designing data collection with human annotators
Python and PyTorch, TensorFlow, or JAX proficiency
Paper at a listed top-tier venue
Apply statistical, engineering, and ethical AI expertise to improve trustworthy, robust LLMs for Cohere agents.