SlipstreamJobsFresh Startup & VC-Backed Jobs

Member of Technical Staff - RL Environments

Cohere - Remote - Remote - posted 2026-08-17

Apply on the company site

SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.

Cohere is building enterprise generative AI models and products. This role focuses on creating reinforcement learning (RL) environments to train and evaluate AI agents for real-world business tasks. You will own the design and implementation of RL environments that simulate enterprise work scenarios. This includes building realistic task environments, generating plausible input data, defining reward structures, and implementing verifiers to measure agent success. You'll train and evaluate agents within these environments, then iterate on both the agents and environments based on performance data and customer feedback. Key responsibilities include: designing new RL environments for different agentic capabilities and industry verticals; training and evaluating agents in those environments; integrating tasks, data, tool implementations, and verifiers; collaborating with modeling and product teams to identify agent performance gaps; working with external vendors to create high-quality environments and building quality assurance tools; automating the discovery of model capability gaps; and systematically measuring agent performance during evaluation and training. You should have deep experience engineering agents optimized for specific use cases, spending significant time analyzing agent trajectories to identify failure points and fix them through model training or harness engineering. You obsess over measuring agentic capabilities and turning measurement into repeatable processes. You've designed reward structures, implemented verifiers, and run annotation workflows to surface insights into agent performance. You've built synthetic data pipelines to scale evaluation and training. Ideally, you use agents yourself daily and have optimized your setup for productivity. Experience with RL training—scaling, troubleshooting, and tuning environments—is a plus. Cohere is remote-friendly with offices in Toronto, London, NYC, San Francisco, Montreal, Paris, Berlin, and Seoul. There is no minimum in-office requirement. The company offers comprehensive benefits including health/dental, retirement matching, 6 weeks paid vacation, parental leave top-up, education stipends, and home office setup credit.

About Cohere

AI / Data / Infrastructure — enterprise generative AI models and tooling for businesses.

Similar roles