Data Scientist, Developer Productivity

About the role

You'll partner with Developer Productivity engineering leadership to define what "developer productivity" means in an AI-first org and to set the strategy for how Anthropic measures, understands, and improves it. This is a space where the playbook doesn't exist yet: AI-assisted development is reshaping how engineers work faster than anyone can measure, and last quarter's answer is already suspect. You'll decide which questions are worth asking, build the evidence to answer them, and stay ready to revise when the ground shifts again.

You'll own the data strategy end-to-end: which metrics earn the org's trust, which investments to push for, which assumptions to challenge — including your own. The space rewards people who hold conclusions loosely, instrument early, and update fast when the data disagrees with the narrative. This role sits at the intersection of data science, developer experience, and frontier AI, with Anthropic's own teams as your users.

Key responsibilities

• Lead ambiguous, high-stakes investigations where the question isn't yet well-formed — from "is Claude making engineers faster?" to "what does 'faster' even mean here?"
• Treat findings as provisional in a space that changes month to month. Bias toward instrumenting first, collecting evidence broadly, and revising the team's priors as the picture sharpens
• Partner with Developer Productivity engineering leadership to set the team's measurement and research agenda — what to study, what to build, what to stop
• Define the metrics framework for developer productivity in an AI-augmented org, and drive its adoption as the basis for tooling and infrastructure investment decisions
• Design and run experiments on internal tooling and workflow changes; build the causal evidence base for what actually moves productivity
• Influence engineering, infrastructure, and product leadership with data. Push back when the data doesn't support the prevailing narrative, and say so plainly when it doesn't support yours either
• Build the analytical foundations (pipelines, dashboards, models) yourself or through partners — staying hands-on and close to the work rather than directing from a distance

Minimum qualifications

• Experience writing production-quality SQL and Python (or a similar language) to build pipelines, dashboards, and models independently
• Experience serving as the primary data or analytics voice in a space where the questions weren't yet well-defined, and helping define them
• A track record of holding conclusions loosely — favoring instrumentation and evidence-gathering over defending a prior position, and revising views in public when the evidence warrants it
• Experience shaping what an engineering or product team worked on, not only measuring what they shipped — being consulted before a decision was made, not just after
• Genuine interest in how AI is changing the way software gets built, with some firsthand experience grappling with the harder, less-defined parts of that question
• Comfort presenting data-backed conclusions to a room of engineers, including when that means saying a built feature isn't moving the needle

Preferred qualifications

• 8+ years of hands-on data science experience, ideally in infrastructure, performance, or platform contexts
• Direct experience with developer productivity, developer experience, or internal tooling, at any scale
• Experience measuring the adoption or impact of AI-assisted workflows, or other tooling where the ground truth was contested
• A track record of building an experimentation or causal-inference practice in an org that didn't already have one
• Prior staff-level or tech-lead scope: setting direction for other ICs and owning a domain's data strategy end to end

Deadline to apply: None. Applications are reviewed on a rolling basis.