Teams at Cohere AI

Evaluate and rank multimodal model outputs, stress-test and document failures, author and edit dataset prompts/responses, build and apply rubrics, perform high-quality annotations across text, image, and structured formats, participate in calibration checks, adapt to evolving tasks, and report model performance trends to cross-functional teams.

Evaluate and label model-generated and human-written code outputs, debug and assess code/agent trajectories, prompt models for complex data science tasks, review visualizations, and report quality trends to improve model behavior.

Evaluate and label model-generated and human-written software engineering outputs for coding, debugging, and codebase-related tasks. Assess agent trajectories, prompt models for complex coding tasks, proofread and improve responses, and report quality and performance trends. Part-time contractor role contributing to model development.