Staff Software Engineer, People Products

Anthropic Anthropic · AI Frontier · United States · Remote · Engineering & Design - Product

Staff Software Engineer focused on building AI-native workflows and LLM-native features for internal people products at Anthropic. The role involves full-stack development, designing and implementing AI tools, evals, and prompts, and working directly with internal stakeholders to iterate quickly. Emphasis on autonomy, shipping products rapidly, and making independent product and architecture decisions in a low-structure environment.

What you'd actually do

  1. Build full-stack end-to-end across the People Products portfolio.
  2. Design and implement AI-native workflows: build tools, evals, prompts, and products. You’ll help define what is possible in applied AI for people processes.
  3. Work directly with internal stakeholders — HR teams, recruiters, managers — to understand problems, gather feedback, and iterate quickly without waiting for requirements to be handed down. No gatekeeping, you are expected to talk to your customers.
  4. Make product and architecture decisions independently in a low-structure environment: knowing when to cut scope, when to ship, and when to ask for input.
  5. Contribute ideas for how the team works, what it builds, and where applied AI can have the most leverage in people workflows.

Skills

Required

  • Full-stack development
  • AI-native workflow design and implementation
  • LLM integration
  • Product and architecture decision making
  • Stakeholder management
  • Independent work

Nice to have

  • Familiarity with MCP (Model Context Protocol)
  • Prior experience building Claude or LLM integrations in production
  • Background at an AI-native company or in a product-focused 0->1 engineering environment
  • Experience with HR tech platforms (Greenhouse, Workday, Rippling)

What the JD emphasized

  • shipped LLM-native features or applications
  • experienced enough to build big features independently
  • move fast without cutting corners
  • engage directly with users and criticism
  • hard feedback

Other signals

  • AI-native workflows
  • LLM-native features or applications
  • prototype to production in days or weeks
  • high autonomy
  • ship constantly