Machine Learning Infrastructure Engineer (Platform)

Physical Intelligence; San Francisco (CA); On-site; USD 155,000 - 210,000; Full time

Job summary

Physical Intelligence in San Francisco, CA is seeking an experienced Infra Platform Engineer to own and evolve AI-native infrastructure. You will operate Kubernetes clusters, design scalable microservice patterns, and support deployments across prototypes to production-grade systems for internal evaluation services and tooling.

You will drive security, observability, and cost visibility, implement IaC (Terraform) and containerization, and collaborate with researchers and engineers to turn

Qualifications

• 4–6 years of infra engineering experience in fast-moving environments.
• Strong knowledge of Kubernetes and cloud platforms (GCP/AWS).
• Experience with Terraform and containerization.
• Ability to design secure, observable, and cost-aware systems.
• Collaborative, ownership-minded with strong communication.

Responsibilities

• Operate and scale Kubernetes clusters and deployment patterns.
• Build a scalable microservice platform for internal tools.
• Drive security, observability, and cost visibility.
• Design authentication/authorization, networking, and quota services.
• Improve developer experience with clear interfaces and workflows.
• Partner with researchers and engineers end-to-end ownership.

Skills

Kubernetes

Cloud platforms (GCP/AWS)

Observability

Distributed systems

Platform engineering

Tools

Terraform

Containerization

Service meshes

Auto scaling

Load balancing

Job description

• The Infrastructure team builds and operates the backbone of everything PI does: from training state-of-the-art VLA models, to orchestrating large-scale simulation, to reliably deploying intelligence across fleets of physical robots
• The team works closely with researchers, robotics runtime, product, and platform engineers to ensure infrastructure scales from prototype to production-grade deployments
• Own and scale AI-native infrastructure: You will operate and evolve Kubernetes clusters and service deployment patterns, and help build a scalable microservice platform for internal systems such as evaluation services, operational tooling, and internal APIs with agent-use as the primary interface. This includes supporting safe rollouts, upgrades, and rollback strategies
• Drive security, observability and cost-aware infrastructure: You will treat security, logging, metrics, tracing, and alerting as first-class platform primitives, and build systems that surface reliability and performance issues early. You will also help improve cost visibility and enable cost-aware decision-making at the infrastructure level
• Harden platform foundations: You will own core infrastructure with multi-cloud considerations, designing authentication/authorization flows, networking architecture, quota and rate-limiting services, and cloud primitives that behave predictably. A major part of this work is reducing infra churn by standardizing patterns, abstractions, and interfaces
• Improve developer experience: You will build clear, documented interfaces for using platform infrastructure, reducing the gap between “I need infra” and “I can run my workload.” This includes supporting consistent local vs. remote development workflows and improving self-serve infrastructure usage
• Collaborate and lead through ownership: You will work closely with researchers and other engineers to understand requirements and constraints, translate fast-moving needs into reusable infrastructure, and own systems end-to-end, from design through operation

Ability to reason about system bottlenecks, performance tuning, and cost optimizations across compute, networking, and storageExperience (4-6 years) working in fast-moving or early-stage environments where ambiguity is normal with demonstrated growth trajectorySolid intuition around scalability, performance, and failure modesDeep experience with cloud platforms (GCP, AWS) and distributed systems: compute orchestration, networking, autoscaling, service meshes, load balancingComfort with Kubernetes, cluster-level reliability, and service-oriented architecturesExperience with infrastructure-as-code (e.g. Terraform), containerization, and modern platform engineering practicesFamiliarity with agent infrastructure, observability, security and sandboxingExtremely strong first-principles thinkingStrong cross-functional communication and ownership mindsetFamiliarity with logging, metrics, tracing, incident response, SLOs, and debugging complex distributed systemsUltimately, we’re looking for someone who can spin up quickly on unfamiliar and ambiguous domains, has a strong sense of ownership, and cares deeply about our mission. Even if you don’t check all the boxes above, we strongly encourage you to apply if this sounds like youExperience with large-scale ML training, evaluation, or simulation infrastructureExperience with secrets management systems (e.g., Doppler)Exposure to robotics, simulation, or real-time systemsBackground in observability, cost optimization, or internal platform tooling