Role Overview
Cohere's Internal Infrastructure team builds and operates the platform for training, evaluating, and serving frontier AI models. This role owns the developer platform—including CI/CD, container images, dev environments, and training instrumentation—used by researchers and ML engineers. You will ensure stability, scalability, and observability of GPU clusters, and participate in a compensated 24x7 on-call rotation.
Responsibilities
- Run secure, scalable developer environments that accelerate researcher velocity.
- Own container image build pipelines and CI/CD tooling.
- Instrument training code and build self-service debugging tooling.
- Establish infrastructure best practices via documentation, AI enablement, and office hours.
- Troubleshoot production failures and performance issues; leave systems easier to operate.
- Build automation, infrastructure-as-code, and observability for maintainability.
Requirements
- Experience running a developer platform in production (CI, image builds, dev environments, internal tooling).
- Proven ability to accelerate ML engineers/researchers without introducing fragility.
- Strong Linux and Kubernetes skills; GPU cluster operations a plus.
- Proficiency in Python or Go; preference for extending solid open-source projects.
- Collaboration with cross-functional teams (researchers, ML engineers) with divergent goals.
- Track record of identifying bottlenecks, proposing fixes, and driving them in fast-paced environments.
Benefits
- Weekly lunch stipend ($75/£75 local equivalent).
- Comprehensive health, dental, and mental health benefits.
- Retirement matching (RRSP, 401K, Pension).
- 100% parental leave top-up for 6 months.
- 6 weeks paid vacation (30 working days).
- Annual enrichment stipends (arts, fitness, workspace, education).
- Remote-friendly with office access in Toronto, London, NYC, SF, Montreal, Paris, Berlin, Seoul; $500 home office stipend.