Applied Research - Evals & Data

Role overview Develop and apply evaluation methods, data practices, and agent systems that improve the real-world performance of advanced AI models. This customer-facing research role connects post-training and reinforcement learning work with deployed workflows, helping teams build agents that are more capable, reliable, and safe. The available description is incomplete, so specific qualifications and the full scope of responsibilities are not clear.

Responsibilities - Design and refine agents for practical workloads, including workflow automation and reasoning-heavy tasks. - Use data from real deployments to improve model policies, reasoning quality, reliability, and safety. - Contribute to infrastructure and evaluation systems for agent development and post-training. - Work with customers on the application of research methods to real use cases.

Benefits and work setup - Full-time position based in New York City.