Technical Program Manager, Science Strategy & Communications

Anthropic•San Francisco, CA

21h•$365,000 - $435,000•Hybrid

About The Position

Anthropic is hiring a Technical Program Manager embedded with its Discovery team to connect researchers advancing Claude’s scientific capabilities with the external scientific community. The role develops and runs programs that help researchers use Claude, including grants, research-group access programs, outreach, and editorial initiatives such as the Anthropic Science Blog. It works closely with Communications, Editorial, Beneficial Deployments, and external partners including universities, funders, and scientific societies. This is a technical, research-oriented position rather than a purely operational role. The successful candidate brings firsthand research experience and an understanding of academia, and will have significant autonomy to make decisions about funding, publishing, and engagement with academic communities after aligning on strategy with Discovery leads. The position also helps shape Anthropic’s external science strategy, gather feedback about researchers’ use of Claude, and translate that feedback into recommendations for research and product teams.

Requirements

PhD in a scientific or mathematical field, with a track record of original research.
Demonstrated writing ability for scientific and general-interest audiences, and sufficient understanding of how modern AI models are trained to write about them accurately.
Experience with the grants process, such as applying for and managing competitive funding; experience designing, reviewing, or administering funding programs is ideal.
Deep engagement with frontier AI, including using these tools in one's own work and understanding their strengths and failures.
Track record of completing multi-stakeholder projects and coordinating people with other high-priority work.
Resourcefulness, high agency, and ability to navigate ambiguity and shifting priorities.
Credibility with researchers and understanding of academic communities, including their incentives, norms, and sensitivities.
Ability to learn quickly, build contextual understanding in unfamiliar scientific domains, and contribute meaningfully to discussions with researchers outside one's field.
Bachelor’s degree or equivalent combination of education, training, and/or experience.
Background in a field relevant to the role, demonstrated through coursework, training, or professional experience.

Nice To Haves

Experience as a program officer, program director, or fellowship lead at a science funder, research agency, or philanthropy.
Published science writing in outlets such as Quanta, Scientific American, or Physics Today, or a substantial public writing presence such as a blog or newsletter.
Research experience using AI or ML methods, or involving the development of scientific software, tools, or datasets.
Experience building partnerships with scientific societies, journals, or research institutions.
Familiarity with AI safety and responsible deployment, including dual-use considerations in fields such as biology.
Experience organizing conferences, workshops, or large coordinated launches.

Responsibilities

Serve as the interface between Anthropic researchers working on Claude’s scientific capabilities and external researchers, translating capability improvements into external use.
Recommend scientific areas and communities to prioritize and how programs, collaborations, and communications should fit together.
Guide the scientific direction of the Anthropic Science Blog; identify and source internal and external stories, draft and edit posts, and maintain a steady cadence of high-quality science writing.
Design, launch, and run external science programs, including AI for Science grants and Team plan offerings for research groups.
Develop partnerships with scientific foundations and grantmaking organizations, and work with Beneficial Deployments on programs focused on high-impact scientific areas.
Identify usage patterns and gather feedback from scientific communities about how researchers use Claude and where capability gaps exist.
Turn feedback from scientific communities into concrete recommendations for research and product teams.
Drive multi-stakeholder projects to completion and coordinate people with competing priorities.