Member of Technical Staff: Data Curriculum Lead

About the Role

You will build the data curriculum—what data our models learn from, in what mix, and in what order—across pre-training and post-training. This is a senior individual contributor role at the heart of model quality, sitting between the data platform and the research teams, and one of our highest-leverage bets.

Responsibilities

• Curriculum Ownership: Build the data curriculum—sourcing, mixing, weighting, sequencing, converting—across pre- and post-training.
• Data-Centric Experiments: Run experiments that quantify how curriculum choices move model quality.
• Tooling: Build tooling that makes curriculum changes fast, measurable, and reproducible.
• Understanding Data: Develop an evidence-based understanding of which data helps, which hurts, and why.
• Cross-Team Influence: Partner across research and data teams and mentor on data-centric methodology.

Requirements

• Data-Centric ML: A track record of improving models through data rather than architecture alone.
• Experimental Rigor: Rigor in experimental design and in correctly attributing model changes to data changes.
• Influence: The ability to lead a cross-cutting work stream and influence multiple teams without owning them.
• Engineering: Sound skills to build the tooling your experiments depend on.
• Communication: Clear communication of results and their implications.

Preferred Qualifications

• Experience with large multimodal or VLA training data.
• Familiarity with data-mixture and scaling-law research.
• A history of data-centric wins on real models.

Benefits

• Competitive total compensation program, including salary, annual cash bonus, and company equity.
• Company-subsidized insurance programs.
• 401(k) with company match.
• Flexible PTO.
• Daily lunch.
• Other benefits.