Data Annotation Specialist, Data Science

Overview

Cohere is hiring a part-time independent contractor for its Data Quality team to support data science and coding evaluation work. This remote contract role is available within Canada and requires a minimum commitment of 16 hours per week. The contract rate is 40 CAD per hour, and the engagement is expected to run for 12 months. Candidates must provide their own laptop.

The work focuses on evaluating coding and data science tasks, reviewing and debugging code, analyzing model trajectories, and assessing how well models handle data visualization and logical flow challenges. The role also involves reviewing machine-written and human-written outputs related to software engineering and data science.

What You'll Do

• Evaluate model responses to coding requests, workflows, and codebase-related questions.
• Assess agent trajectories and model performance on code generation, debugging, and tabular or graphic manipulation tasks.
• Prompt models with complex data science problems and review the quality and accuracy of the outputs.
• Label, proofread, and improve software engineering-related content produced by both humans and models.
• Track and report quality and performance trends for model behavior and project assignments.

Requirements

• At least 3 years of relevant industry experience working on real-world data science problems and pipelines.
• Strong data analysis and visualization skills.
• Proficiency in Python and common data science libraries such as numpy, pandas, matplotlib, sqlite, or similar tools.
• Strong understanding of SQL and experience working with data and file formats such as markdown, JSON, XML, YAML, and HTML.
• Experience rewriting, proofreading, and giving feedback on code.
• Ability to work as an independent contractor and maintain IP confidentiality.

Preferred Qualifications

• Familiarity with code agents such as OpenCode, Claude Code, or Codex.
• Experience evaluating agent trajectories.