Member of Technical Staff – Infrastructure

Description

The company is a frontier lab working on action and world models that perceive, predict and act across virtual and physical environments.

It is closely tied to the company, the largest and fastest-growing platform for gaming clips, which millions of gamers use each year to record, share and find games.

Since October 2025 the company has raised over $650M from Khosla, GC, Valor and Point72, and recently closed a round at a $6.2B valuation.

Responsibilities

• Train action policies and world models on the biggest action-labeled gaming dataset available, and run the company platform, which millions of gamers use daily.
• Both depend on one infrastructure: billions of gameplay clips a year must be stored, moved and served, handling research workloads one day and consumer traffic the next.
• Own infrastructure end to end as a generalist, centered on cloud and storage: the cloud architecture and the fleet on it, object storage for petabytes of gameplay video, and the data pipelines and CI/CD around them.
• Make architecture decisions before scale turns them into emergencies, and anticipate where current systems will fail at 10x the data, traffic or team size.
• Write application code when the problem calls for it, moving from infrastructure design directly into building the software.
• Build the infrastructure function in New York: set technical direction, hire and grow the team, and define how systems are designed, reviewed, deployed and run, while staying hands-on.
• Work with research and product engineering across the company and the company, and serve as the go-to person for system design.

Requirements

• Deep hands-on experience with cloud and storage in production on GCP and CoreWeave, including owning Terraform, IAM and Kubernetes on a major cloud.
• Experience making placement, tiering and egress cost decisions for object storage at scale.
• 6+ years of hands-on work with planet-scale distributed systems such as search infrastructure, large storage or query systems, or time series systems, with solid grounding in storage, query engines and consistency, and in running them in production.
• A record of building infrastructure from scratch rather than only operating what others designed, ideally as a founder or one of the first infrastructure engineers, and a history of choosing hard problems.
• Experience leading infrastructure while still writing code: setting technical direction, raising the bar for the engineers around you, ideally including hiring them.
• Range from low-level infrastructure up to application code, so you do not need a handoff to build what your infrastructure supports.
• High agency and sound judgment under ambiguity, since priorities shift between a research lab and a live consumer product. You identify what matters and act on it.
• Petabyte-scale video or media experience, or real games background and familiarity with gaming data.
• Depth in ML tooling such as PyTorch or JAX and world models, in data orchestration such as Dagster, in GPU training clusters, or in systems work in Rust or C++, including substantial open source contributions.
• Experience running infrastructure that serves research workloads and consumer production traffic at the same time.

Conditions

The infrastructure team is based in New York.