Senior Infrastructure Engineer — Build AI at Scale

Cognition; New York (NY); On-site; USD 260,000 - 300,000; Full time

Benefits offered by this job

Private chef

Cozy slippers

Endless snacks

Job summary

Cognition, an applied AI lab, builds end-to-end software agents including Devin and Windsurf. We pursue reliable, scalable infrastructure to run long-horizon tasks, sandboxed environments, and subagents. You will own compute, orchestration, networking, and platform systems central to our products.

As a member of a tightly knit team from Scale AI, Palantir, and Google DeepMind, you’ll shape the backbone that enables product teams to ship rapidly with strong reliability and security.

Qualifications

• Experience designing and operating large-scale distributed infrastructure.
• Hands-on proficiency with Kubernetes, cloud platforms (AWS, GCP, or Azure), and Terraform.
• Strong software engineering fundamentals and Python experience.
• Observability and alerting—meaningful dashboards and metrics.
• Security, sandboxing, and multi-tenant compute expertise.
• Prior frontier AI or developer tools experience is a plus.
• BS, MS or equivalent from a top-tier university.

Responsibilities

• Own agent execution infrastructure and sandboxed compute environments at scale.
• Build and maintain the internal developer platform (CI/CD, deployment tooling).
• Drive reliability, define SLOs, and lead incident response and postmortems.
• Scale with product growth by forecasting capacity and architecture needs.
• Collaborate with product engineers and researchers to meet requirements.

Skills

Deep systems engineering

Cloud and container expertise

Software engineering fundamentals

Observability instincts

Security and isolation mindset

Relevant industry experience

Top-tier university degree

Education

BS/MS in CS/Engineering

Tools

Kubernetes

AWS

GCP

Azure

Terraform

Job description

Cognition, an applied AI lab, builds end-to-end software agents including Devin and Windsurf. We pursue reliable, scalable infrastructure to run long-horizon tasks, sandboxed environments, and subagents. You will own compute, orchestration, networking, and platform systems central to our products.

As a member of a tightly knit team from Scale AI, Palantir, and Google DeepMind, you’ll shape the backbone that enables product teams to ship rapidly with strong reliability and security.