Technical Lead Manager - Training Runtime, Data(set) Movement

About the Team

Training Runtime builds the distributed systems that power OpenAI's largest model training runs - most recently GPT-5.5! The Data Movement area owns the infrastructure that keeps training jobs supplied with the right data at the right time, and keeps model state moving safely and efficiently across large clusters.

Our work spans machine learning systems, distributed storage, high-throughput data loading, reliability engineering, and developer experience. Success means researchers can move quickly while training runs remain fast, reproducible, debuggable, and resilient at scale.

About the Role

We are looking for a deeply hands-on Technical Lead Manager to own datasets throughout our training infrastructure. This person will set the direction for how training jobs read data: the APIs, storage contracts, versioning model, benchmarks, debugging tools, and reliability guarantees that make data access consistent across current and future training frameworks.

You will begin as the primary technical owner for dataset reads, working directly in the code while aligning researchers, training framework owners, storage teams, and infrastructure partners around a durable platform. The problem is deceptively hard at frontier scale: make enormous, heterogeneous datasets easy to consume, correct across distributed workers, observable when something goes wrong, and flexible enough to support pretraining, reinforcement learning, and multimodal training.

In this role, you will

Design and build a unified dataset read platform for multiple current and future training frameworks.

Define dataset APIs, storage-format expectations, registration/versioning, and migration paths that make data access reproducible and maintainable.

Build reliability into the read path, including stateful iteration, caching, fast restart, recovery, and clear operational contracts.

Build terminal and web-based visualizers that let teams inspect text, multimodal, and reinforcement learning data late in the pipeline, where bugs are most visible.

Write and review production code in core data loading, service, caching, and reliability paths.

Partner with teams working on training frameworks, reinforcement learning, multimodal models, storage, runtime, and cluster infrastructure.

Over Time

The long-term goal is a team that owns fast, correct, scalable, and reliable in-cluster data movement for training: data that comes in, data that goes out, and data that moves around inside the cluster. After ramping on datasets, this role will expand to TLM ownership for broader data movement systems, including checkpoint loads/saves and snapshot transfers, while partnering closely with existing technical leads and adjacent infrastructure teams.

You might thrive in this role if you:

Have built or owned dataset, data loading, storage, or distributed training infrastructure at large scale (e.g. torch.utils.data )

Care equally about API design, debugging ergonomics, performance, and bit-level correctness.

Understand the failure modes of large distributed training jobs and know how data systems can create or prevent them.

Have experience with stateful iterators, checkpoint/restart semantics, caching, remote services, or high-throughput storage reads.

Are comfortable working across Python and lower-level systems code; Rust or C++ experience is useful but not required.

Have worked with multimodal, video, reinforcement learning, or pretraining data pipelines where small data bugs are expensive and hard to diagnose.

Can lead through code and technical judgment before a team exists, and can later manage engineers without losing the hands-on edge.

Obsess over developer experience by eliminating friction, such as manual preprocessing scripts and niche cluster-specific bugs, ensuring a reliable and efficient experience for researchers.