Software Engineer, Staff: Applied AI, Science & Engineering

Anthropic is building safe, reliable, and interpretable AI systems that benefit humanity. The company's Claude model has demonstrated remarkable capabilities in research and analysis, yet the hardest scientific and engineering problems remain unsolved because they exist within complex real-world environments with proprietary data, specialized tools, and domain constraints. You'll join the Applied AI, Science & Engineering team, which bridges the gap between Claude's capabilities and the practical needs of researchers and engineers tackling challenges in energy, materials, hardware design, and related physical-world fields.

Anthropic operates as a research-driven AI safety company founded by former members of OpenAI, with a focus on developing AI systems that are interpretable and aligned with human values. The company has grown rapidly while maintaining its commitment to responsible AI development, and this role reflects that growth into applied domains where AI can demonstrably solve real problems.

What you'll do

• Develop agent frameworks and research loops that enable Claude to work through entire problem-solving workflows, from literature review and hypothesis generation through simulation, analysis, and iterative design

• Integrate scientific computing tools including finite element analysis, computational fluid dynamics, electromagnetic modeling, and molecular simulation software so Claude can operate them reliably at scale

• Create robust evaluations that measure whether Claude's scientific work is accurate, and communicate honestly about performance gaps and limitations

• Partner directly with scientists and engineers at client organizations to understand their workflows, constraints, data, and tools, translating open-ended research questions into concrete, verifiable tasks with measurable success criteria

• Build and test pilot systems with partners, rapidly iterating based on real-world feedback and removing anything that doesn't meaningfully advance their work

• Generalize successful approaches from individual engagements into reusable tools, shared components, and documented processes that accelerate future projects

• Serve as the technical expert explaining how Claude performs on scientific and engineering tasks, articulating capabilities, limitations, and tradeoffs to both technical and non-technical audiences

• Work closely with Anthropic's research and product teams to identify where models fall short and inform the development of future capabilities

What you'll bring

• At least eight years of professional software engineering experience with strong fundamentals across the full stack, from data pipelines through tooling to user-facing interfaces

• Proven experience building production systems with large language models, including work with agents, tool integration, or evaluation frameworks

• Background tackling technical problems in physics, materials science, chemistry, electrical engineering, mechanical engineering, or related scientific fields

• Demonstrated success with zero-to-one initiatives in startup or startup-like settings, comfortable with ambiguity and rapid learning

• Ability to engage deeply with expert users, understand their workflows and constraints, and maintain focus on building the right system rather than simply delivering requested features

• Strong judgment about what can be verified and validated, with willingness to reject solutions that don't meet quality standards

• High agency and self-direction, capacity to master new domains quickly, and ability to hold informed opinions while remaining intellectually flexible

• Clear communication skills across technical researchers, engineers, and external partners, with genuine consideration for the societal implications of your work

Nice to have

• Advanced degree or research experience in a physical science or engineering discipline

• Direct experience with scientific computing, numerical methods, optimization algorithms, or simulation software

• Background in industrial R&D, national laboratories, or deep-tech companies working in energy, materials, manufacturing, or hardware

• Prior work embedded with customers or technical partners in forward-deployed, applied AI, or solutions engineering roles where you owned the resulting software

• Experience in ML research, reinforcement learning environments, or evaluation design

What they offer

• Annual salary range of 405,000 to 485,000 USD

• Full-time employment based in New York City, San Francisco, or Seattle