RE/RS, Data Understanding - Foundations

OpenAI, an AI research and deployment organization dedicated to ensuring that general‑purpose artificial intelligence benefits all of humanity, is looking for a researcher to join its Data Understanding team. This team builds the high‑quality datasets and their quantized representations that power OpenAI's large‑scale model training, handling everything from data synthesis and filtering to deduplication, quality control and tokenization. In this role you will treat data quality and curation as core research challenges. Your daily work will involve inventing new techniques for selecting, merging and transforming data, constructing datasets that expand model capabilities, and designing rigorous experiments that reveal how data choices and interventions influence learning and downstream behavior. You will collaborate closely with frontier‑model teams and work with web‑scale data to generate evidence about which approaches succeed, then turn those findings into scalable data‑processing pipelines that can be used in massive training runs. The position requires a proven track record of delivering new or improved machine‑learning ideas, demonstrated through publications, projects or applied research. You must be capable of defining and steering a research agenda, from identifying the right problems to seeing long‑term work through to measurable impact. A genuine enthusiasm for OpenAI's empirical, collaborative research culture is also essential. Preferred, though not mandatory, qualities include a thoughtful stance on AI's broader impact, particularly issues of privacy, provenance and data quality, and hands‑on experience building high‑performance deep‑learning or large‑scale data‑processing systems. OpenAI values diverse perspectives and offers an inclusive environment that supports reasonable accommodations for applicants with disabilities. The role is based in the United States, with standard security and confidentiality obligations attached to the handling of proprietary data and hardware. What the role asks for: - Proven record of novel ML ideas via publications or projects - Ability to own and drive a research agenda - Enthusiasm for OpenAI's empirical, collaborative research style - Nice-to-have: awareness of AI impact, privacy, provenance, data quality - Nice-to-have: experience building high‑performance deep‑learning or large‑scale data pipelines