Temporary Biotech Research Scientist Jobs (NOW HIRING)

We're seeking an exceptional Research Scientist to join our Life Sciences team at Anthropic. Our team is building a world-class research group focused on making Claude a superhuman life sciences research assistant. This role sits at the intersection of machine learning, software engineering, and biology - you'll directly improve model capabilities on scientific tasks through post-training, evaluation design, and RL environment development.

As a core member of our Life Sciences team, you'll work in a high-impact team that translates deep biological domain knowledge into model training objectives, benchmarks, and agentic workflows. You'll help establish Anthropic as a leader in AI-accelerated biology while shaping how frontier models reason about and execute computational biology tasks.

This role offers a unique opportunity to shape how frontier AI models learn to do biology. You'll work alongside some of the world's best AI researchers while tackling problems that matter for human health and scientific understanding. If you're excited about turning your computational biology expertise into model capabilities, we want to hear from you.

• Build and ship agentic tools and integrations that let Claude execute real life science workflows - bioinformatics pipelines, database queries, analysis notebooks, literature review

• Design and build evaluation benchmarks that measure model capabilities on biology tasks - figure interpretation, bioinformatics, protocol reasoning, literature synthesis

• Work closely with product and design teams to scope, prototype, and ship features for life sciences users

• Partner with external biotech, pharma, and academic users to understand their workflows and turn feedback into product improvements

• Build and maintain the engineering infrastructure behind our biology product surface - tool scaffolding, data pipelines, eval harnesses

• Translate biological domain knowledge into product requirements and evaluation criteria that guide model improvement

• Experience applying ML and software engineering to biological problems - computational biology, bioinformatics, protein ML, genomics, or similar

• Experience working in drug discovery or development at a biotech or pharma company, or conducted fundamental research in an academic setting - with an understanding of what real scientific workflows look like and where they break down

• Strong software engineering skills: comfortable building production-quality Python, working in large codebases, and owning infrastructure end-to-end

• Hands-on experience training or fine-tuning ML models (LLMs, protein language models, or other deep learning architectures)

• A track record of shipping computational tools or pipelines that biologists actually use

• Comfortable navigating ambiguity and defining problems in a rapidly evolving research environment

• Able to work independently while collaborating tightly with research, product, and domain-expert teams

• Results-oriented with a bias toward rapid iteration and measurable impact

• Passionate about using AI to accelerate scientific discovery while maintaining high ethical standards

• 5+ years of experience applying ML and software engineering to biological problems - computational biology, bioinformatics, protein ML, genomics, or similar

• Ph.D. in computational biology, bioinformatics, bioengineering, CS, or a related quantitative field - or equivalent industry experience

• Experience with LLM post-training: RLHF, RL from verifiable rewards, SFT data curation, or eval-driven development

• Direct experience with therapeutic discovery pipelines - target identification, lead optimization, ADMET modeling, or clinical data analysis

• Familiarity with bioinformatics tooling and pipelines (sequence analysis, structure prediction, single-cell, variant calling, etc.)

• Experience building agentic systems or tool-use environments

• Published research in ML for biology, or open-source contributions to computational biology tools

• Fluency with biological databases (UniProt, PDB, Ensembl, NCBI) and the ability to reason about their schemas and failure modes