Product Manager, Claude Science

Anthropic•Seattle, WA

14h•$305,000 - $385,000•Hybrid

About The Position

Anthropic's mission is to create reliable, interpretable, and steerable AI systems. We believe powerful AI can compress a century of scientific progress into a decade. Claude Science is a big part of building toward that future. Claude Science is an AI workbench that gives researchers a single environment for work that today spans dozens of disconnected tools — literature, specialized databases, scientific computing, analysis, and publication-ready outputs. Claude Science already renders protein structures, genome tracks, and chemical structures natively, and coordinates multi-agent workflows with built-in review for citation and calculation errors — and we're just getting started. Much of this is being built 0→1 right now, in a category no one has defined yet. We're growing the Claude Science product team. As a Product Manager on Claude Science, you will own a major area of the roadmap end to end — think expanding into new scientific fields, landing new model capabilities with researchers, or making the workbench ready for enterprise R&D teams — and help shape the overall product direction alongside the rest of the team. You'll spend real time with researchers in academic labs, biotech and pharma R&D, and research institutes to understand how science actually gets done — from hypothesis to analysis to publication and beyond— and turn what you learn into shipped product. You'll partner directly with our research teams to make the models better at science, help decide which fields and workflows we go after next, and shape what it takes for Claude Science to be trusted in enterprise and regulated research environments. We're looking for experienced product managers and founders with real scientific or technical depth: people who can hold their own in a conversation with a structural biologist or a computational chemist, get in the weeds of model transcripts and evals, and still zoom out to help define a product category. You thrive in ambiguity, uncover non-obvious use cases for nascent capabilities, and bring those insights back to research. Our north star is for Claude Science to accelerate scientific progress — across biology, chemistry, physics, and beyond — from early discovery through real-world application, by an order of magnitude.

Requirements

Have product management experience shipping technical products in close partnership with engineering and design, or equivalent experience driving product direction as a founder, engineer, or scientist
Have a scientific or deeply technical background — enough to reason with domain experts about their workflows and to diagnose where and why a model is falling short on a scientific task
Have a strong grasp of AI and LLM concepts, use AI tools daily, and are comfortable going deep on model behavior, prompting, and evaluation methodology
Are data-driven: you ground prioritization in usage data, eval results, and evidence from users rather than intuition alone
Have strong user empathy and can synthesize vague or contradictory feedback from expert users into actionable priorities
Navigate and execute amid ambiguity as a first-principles thinker, flexing across scientific domains based on the problem at hand and finding simple, easy-to-understand solutions
Communicate clearly in writing and can align research, engineering, and go-to-market stakeholders through influence rather than authority
Think carefully and creatively about the risks and benefits of putting frontier scientific capabilities into the world, beyond past checklists and playbooks

Nice To Haves

8+ years in product management or equivalent product leadership, including taking products from 0 to 1 and scaling them; founder experience is a plus
An advanced degree or hands-on research experience in biology, chemistry, physics, materials science, or a related field — or experience building software for scientists (e.g., computational biology, cheminformatics, lab informatics, simulation, or scientific data platforms)
Have personally built evals or benchmarks for model capabilities, ideally agentic or scientific ones
Experience shipping products into life sciences, pharma, or other regulated research environments, including familiarity with their data governance and compliance expectations
A track record of launching ambitious products that found distribution or commercial success
Experience designing trusted-access or staged-rollout programs for sensitive capabilities
A creative, hacker spirit and a love of solving hard puzzles

Responsibilities

Own vision, strategy, roadmap, and execution for a major area of Claude Science, and contribute to overall product direction in a category that is still being defined
Spend meaningful time with working scientists — in interviews, working sessions, and key customer conversations — and synthesize what you learn into clear priorities, requirements, and success metrics
Partner with research to make Claude better at science: define target behaviors, build and shape evals grounded in real scientific workflows, surface failure modes from real usage, and feed them back into model development
Drive model and feature launches in your area end to end: define readiness criteria, coordinate across research, engineering, design, and go-to-market, run early access programs, and make sure launches land well with researchers
Identify which scientific domains, data sources, tools, and integrations Claude Science should support next, and build the case with evidence from users, usage data, and the broader landscape
Prototype ideas yourself with Claude to validate them before committing engineering time
Work with safeguards, policy, security, and go-to-market partners so that powerful scientific capabilities reach trusted researchers responsibly, including in enterprise and regulated research settings
Drive enterprise readiness — security and compliance reviews, admin and deployment controls, and the procurement realities of pharma and biotech — so large organizations can adopt it with confidence
Prioritize ruthlessly across a broad surface — scientific domains, customer segments, and competing workflows — and make clear calls on what our products serve deeply first, holding the line on MVP versus ideal state
Define how success is measured for AI research and development products — adoption and retention among working scientists and operators, time-to-result on real analyses, trust in outputs — and use those measures to steer planning
Maintain an objective, current view of the AI-for-science ecosystem — models, tools, benchmarks, and competitors — and of where Claude Science stands within it