Senior Research Engineer - Safety Tooling and Data

About the role

This role focuses on developing the core tooling and infrastructure necessary for creating safer AI models. You will design and implement data pipelines to facilitate frequent data generation and annotation, ensuring these systems integrate smoothly with ongoing model experimentation. The goal is to establish robust, standardized processes for data validation, analysis, and continuous improvement for both training and evaluation datasets.

The work involves creating cohesive data infrastructure that operates in parallel with model development. You will collaborate closely with the ML modeling team to ensure data capabilities meet their experimental needs. A key aspect is to develop systematic analysis frameworks to identify new data sources and pinpoint coverage gaps, ultimately contributing to the capabilities of Cohere's foundation AI models.

Success in this position means delivering well-documented, maintainable solutions that become team standards. You will be responsible for owning complex technical projects from initial concept through to deployment, applying strong software engineering principles and an opinionated approach to technical architecture to make sound decisions that advance the safety and performance of AI systems.

The salary for this role ranges from $230,000 to $380,000 USD.

Skills that matter here

• Python: This is a primary language for implementing data pipeline tooling and analysis frameworks.
• PyTorch: Experience with ML frameworks like PyTorch is required, indicating work at the intersection of data and model development.
• BigQuery: Proficiency in big data analytics tools like BigQuery is essential for managing and processing large datasets.
• SQL: SQL skills are necessary for data querying, analysis, and managing data within the infrastructure.

Who this role suits

• A person who thrives on owning complex technical projects from start to finish.
• Someone with a strong background in both software engineering and statistical analysis.
• An individual who prefers to establish and maintain high standards for technical architecture and documentation.
• A collaborator who can align data capabilities directly with machine learning experimental needs.

From the employer

• Design and implement robust data pipeline tooling that enables frequent, low-friction data generation and annotation
• Create cohesive data infrastructure that supports continuous parallel operation with model experimentation
• Establish standardized processes for data validation, analysis, and improvement of data both training and evaluation
• Collaborate with the ML modeling team to align data capabilities with experimental needs
• Maintain opinionated, well-documented solutions that become team standards
• Develop systematic analysis frameworks to identify incoming data sources and benchmark coverage gaps

• Extremely strong software engineering skills.
• Strong statistical skills and experience evaluating scientific experiments related to data collection and model performance.
• Proficiency in programming languages such as Python and ML frameworks (e.g., PyTorch) and Big Data Analytics (e.g. BigQuery, SQL)
• Demonstrated ability to own complex technical projects from conception to deployment
• Opinionated approach to technical architecture with ability to make principled decisions
• Understanding of ML data requirements and the intersection of data engineering with modeling workflows

• A weekly lunch stipend of $75/£75 or equivalent in your local currency for lunch.
• Full health and dental benefits, including a separate budget for mental health.
• RRSP matching, 401K, Pension Scheme.
• 100% Parental Leave top-up for up to 6 months, for either parent.
• Annual enrichment benefits:
• Arts & culture, fitness/wellness, quality time, and a workspace improvement credit.
• Education & learning stipend for conferences, courses, and coaching.
• 6 weeks of paid vacation (30 working days!)
• Budget for traveling to other offices if you are remote, plus an annual company offsite.