Technical Program Manager, Storage & Data Infrastructure
OpenAI
- Pay
- $257k–$445k
- Workplace
- On-site
- Location
- San Francisco · California · United States
- First seen
- 2 days ago
- Last seen
- 4 hours ago
- Board
- Ashby
Summary
About the Team OpenAI's data and storage infrastructure spans data platforms, online databases, and file/object storage.
Posting
About the Team OpenAI's data and storage infrastructure spans data platforms, online databases, and file/object storage. These systems underpin data ingestion and processing, durable persistence, indexing and retrieval, and product file experiences. As frontier models and agents evolve how they use memory, history and snapshots, the underlying architecture increasingly shapes the capabilities products can deliver—and their latency, reliability, cost and efficiency. About the Role We are looking for a technically deep TPM to independently define and lead multiple programs across data platforms, online databases and storage infrastructure. You will connect model, product and data-consumer requirements to architecture, and work with the relevant engineering teams to take new capabilities through production adoption and repeatable expansion. The design scope is exabyte-scale storage and infrastructure spanning multiple millions of CPU cores. The challenge is not simply forecasting more resources: it is making complete, workload-ready capacity repeatable, with a clear path from product requirements through architecture, deployment and validation. A data pipeline, database query, file operation or execution snapshot can affect whether a product or agent succeeds; you will connect those outcomes to the systems underneath. This role is based in San Francisco, CA. We use a hybrid work model of 3 days in the office per week and offer relocation assistance to new employees. In this role, you will: - Translate model, product and data-platform needs into precise access patterns, consistency, durability, freshness, availability and scalability requirements. Connect memory, history, retrieval and resumable work to capability and end-to-end latency. - Partner with engineering to transform data and storage architecture into repeatable scale units: standardized provisioning, placement, routing, data movement and readiness checks that bring storage, compute and networking online together. Tie each expansion to the workloads it can serve. - Lead cross-stack programs connecting ingestion and processing, databases and indexes, and file/object storage. Make data ownership, schema compatibility, change-data-capture, replay and consumer-readiness contracts explicit so the full data path remains correct and usable. - Make cost and efficiency architectural inputs. Evaluate physical versus logical footprint, index and replication amplification, redundant copies, tiering