Claude Science creator Alec Tarashansky joins OpenEvidence

Why it matters

Tarashansky's move connects a research-workflow builder with a fast-growing clinical evidence platform, where traceable answers and expert oversight are central product requirements.

Alec Tarashansky (@AlecTPhD), who helped lead Anthropic's Claude Science workbench, said on October 3rd that he has joined OpenEvidence to work on AI for medicine. His move takes a builder of tools for scientific research into a company focused on helping doctors find and use medical evidence.

Tarashansky framed the move around a shared problem: making expert knowledge more usable without asking people to do all the information-processing work themselves. In his announcement on X, he described Claude Science as a product built so that "scientists' curiosity, not their capacity to execute" sets the limit on discovery. At OpenEvidence, he wrote, he hopes to bring that approach to doctors at the point of care.

From research workflows to clinical evidence

The work Tarashansky leaves behind is a broad research environment, not a single-purpose chatbot. Anthropic introduced Claude Science on June 30th as a beta app that connects Claude to more than 60 scientific skills and data sources, can work with local or remote computing resources, and keeps code and other artifacts alongside generated analyses. Researchers can use it for tasks including literature analysis, genomics and protein-structure work.

Anthropic named Tarashansky as Claude Science's lead engineer at an August 20th product event. His own description of having created the product is consistent with that public role, while Anthropic's launch materials describe a larger system and team rather than a solo build.

The product's design puts provenance and review alongside automation: Claude Science says it preserves the code, environment and conversation history behind figures, and includes an agent that checks citations and calculations. Anthropic has also published examples of researchers using the workbench for tasks such as single-cell analysis and long-form literature reviews. RuntimeWire recently covered Anthropic's BootLoops project for exact calculations, another effort built around applying AI to real scientific workflows while keeping experts involved in deciding what results matter.

OpenEvidence addresses a narrower but higher-stakes setting. The company's product is an AI medical-search and clinical decision-support tool for verified physicians, designed to return answers grounded in medical literature and guidance. A research environment helps scientists move from data and tools toward analysis; a clinical evidence product has to help doctors retrieve and assess information relevant to patient care. Tarashansky's post describes his interest in carrying a workflow-building approach across that divide, rather than announcing a specific new product or feature.

A well-funded destination, with the job details still his to define

OpenEvidence has attracted substantial investor backing as it builds in that physician-focused market. On January 21st, the company announced a $250 million Series D at a $12 billion valuation, co-led by Thrive Capital and DST Global, according to TechCrunch's report. The round gives useful context for the scale of the business Tarashansky is joining; it does not disclose his role, compensation or remit.

His prior path helps explain the combination of science and software in his work. Tarashansky spent three and a half years at Chan Zuckerberg Initiative before joining Anthropic as a software engineer on its life sciences team, he wrote in a post about that move. Earlier, he pursued bioengineering research at Stanford, where the university listed him as a Bio-X fellow working on stem-cell lineage competition in tissue. He arrived at Anthropic with experience spanning computational biology and software, a combination reflected in Claude Science's attempt to connect research tools, data and compute in one environment.

The change of employer does not establish that Anthropic is backing away from science. In August, Anthropic announced discounted or free Claude seats for 10,000 scientists, and in September it announced a new life sciences research group and laboratory. Both announcements predate Tarashansky's October 3rd move and point to continued investment in the field even as a product builder takes his experience elsewhere.

For OpenEvidence, the hire brings someone who has worked on a system for reducing friction across a complicated expert workflow. The question is how that experience translates into a medical product where doctors need evidence they can inspect and apply, not simply faster answers. Tarashansky's announcement states the ambition; it does not yet specify what he will build or when it will reach users.