We're looking for someone to lead developer experience and data tooling for our pre-training data team. This person will build internal tools and infrastructure that make the team more productive—dashboards, CLIs, data exploration UIs, and the systems that tie them together.
The role is focused on DX and tooling—we're looking for someone who genuinely loves this work.
What You'll Do
Lead tooling efforts across the stack: build systems, continuous integration, CLI tools, and internal web UIs
Build internal tools for exploring datasets, labeling data, reviewing data quality, and tracking data inventory
Improve ergonomics of data infrastructure—IO patterns in Ray/dataflow jobs, dataset tracking, pipeline observability
Identify opportunities by engaging with the team, listening to pain points, and proactively improving workflows
Raise the bar on code organization, packaging, and engineering best practices
What We're Looking For
Must-Haves
Strong software engineering fundamentals
Genuine care for developer experience and best practices in code organization
Good communicator who engages with teammates to understand their needs
Bias toward action—sees something broken and fixes it
Based in San Francisco (this role is in-office)
Ideal Background (in rough priority order)
Open source contributor — someone in the mold of tools like Ruff, uv, or similar developer-facing projects
Build systems / CI experience — has written or maintained build systems, CI pipelines, or developer tooling at scale
Startup product dev — comfortable moving fast, shipping throwaway prototypes, iterating quickly
Data infrastructure experience — familiarity with OLAP engines, columnar storage formats, or similar
Not Required
Deep ML/AI expertise (this is a tooling role, not a modeling role)
Prior experience specifically in "data engineering" pipelines—we care more about tooling instincts than domain experience
Why This Role
You'll have significant ownership over how a high-performing team works day-to-day. The scope is broad, the feedback loops are fast, and the work directly impacts how quickly we can move on core research and data efforts.