ML Data Infrastructure Engineer

Physical Intelligence; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Physical Intelligence seeks an ML Infra Engineer (Data Systems) to build and operate data infrastructure enabling large-scale robot learning. You’ll design high-throughput ingestion, batch and streaming pipelines, and scalable storage to support training and evaluation.

You will optimize dataloaders, caching, and metadata systems while collaborating with researchers and roboticists to translate evolving data needs into robust, reliable systems.

Qualifications

• Strong software engineering fundamentals.
• Experience building distributed systems or large-scale data pipelines.
• Comfort reasoning about performance, memory, I/O, and storage efficiency.
• Familiarity with batch and/or streaming processing systems.
• Experience with object storage systems and data format tradeoffs.
• Ownership mindset: design, build, operate, and iterate on systems end-to-end.
• Enjoy working closely with researchers and unblocking fast-moving projects.

Responsibilities

• Design and implement high-throughput data ingestion and processing pipelines for multimodal data.
• Operate large-scale batch and streaming workflows across massive datasets.
• Design storage layouts and metadata systems for scalable access patterns.
• Build data lifecycle tooling including backfills, dataset rebuilds, and garbage collection.
• Optimize dataloaders, sharding, prefetching, and caching to shorten the data-to-training path.
• Develop scalable metadata stores for datasets and training artifacts.
• Move petabytes of data efficiently across clusters and environments.
• Enhance observability, validation, and guardrails to prevent data regressions.
• Collaborate with researchers and engineers to translate data needs into robust systems.

Skills

Distributed systems

Data pipelines

Performance optimization

Memory management

Batch processing

Streaming systems

Observability

Tools

ClickHouse

Ray

Flink

Spark

Job description

Physical Intelligence seeks an ML Infra Engineer (Data Systems) to build and operate data infrastructure enabling large-scale robot learning. You’ll design high-throughput ingestion, batch and streaming pipelines, and scalable storage to support training and evaluation.

You will optimize dataloaders, caching, and metadata systems while collaborating with researchers and roboticists to translate evolving data needs into robust, reliable systems.