Research Engineer, ML Infrastructure

About the role

cognition is an applied AI lab building end‑to‑end software agents, including Devin, the first AI software engineer. We are looking for a Research Engineer to own and evolve the core ML infrastructure that powers our research, ensuring fast, reliable, and scalable training pipelines.

Key responsibilities

• Design, implement, and maintain distributed training infrastructure across large GPU clusters.
• Build and operate the systems that run massive agent rollouts in VM sandboxes at RL scale.
• Profile and optimize end‑to‑end training throughput, addressing bottlenecks in data loading, communication, and compute.
• Develop experiment orchestration tools that streamline launch, tracking, and analysis of research runs.
• Engineer high‑throughput data pipelines with strong reproducibility and quality guarantees.
• Diagnose and resolve failures across GPUs, networking, numerics, and data, improving system reliability.

Required profile

• Strong understanding of large‑scale machine‑learning systems and research workflows.
• Experience working directly with researchers to anticipate infrastructure needs.
• Ability to build production‑grade systems that handle thousands of GPUs simultaneously.

Required skills

• Distributed training
• GPU computing
• Data pipeline engineering
• Experiment orchestration

What we offer

• Opportunity to work with a talent‑dense team from leading AI companies.
• Cutting‑edge problems at the intersection of research and engineering.
• Competitive compensation and equity in a fast‑growing AI startup.