Software Engineer, Memory & Search

Software Engineer, Memory & Search

Anthropic develops reliable, interpretable, and steerable AI systems designed to be safe and beneficial. The company brings together researchers, engineers, policy experts, and business leaders to advance the field of beneficial AI. As the AI field evolves rapidly, Anthropic is expanding its teams to tackle complex technical and product challenges.

The Memory & Search team at Anthropic owns a critical part of Claude's user experience: how Claude remembers conversations and interactions, and how it retrieves relevant information when needed. This spans everything from what users see on Claude.ai and Cowork to how Claude Code surfaces past work. The team recently transitioned from periodic memory summaries to a dynamic, editable system that follows users across all products. Search capabilities evolved from a major user pain point to a trusted feature. Now the focus is discovering which product experiences make memory and search most compelling and integral to using Claude.

You'll join as a product engineer who can ship user-facing features at scale while maintaining strong product instincts. The Memory & Search team has deep infrastructure expertise and is seeking engineers who've built consumer or startup products, understand what makes software feel good to use, and can think through how models, tools, and interfaces work together in agentic systems. You'll run quick, demo-driven experiments, measure real user behavior, and decide what to develop further based on evidence.

What you'll do

• Design and deliver end-to-end memory and search features across Claude.ai, Cowork and Claude Code, from backend services through model tooling to the user interface

• Conduct product experiments by forming hypotheses, building prototypes, testing with real users, analyzing results, and making clear decisions about what to pursue

• Influence how Claude retrieves and uses information in its agentic reasoning loop, including deciding what to store, when to search, how to rank results, and keeping users in control

• Define success metrics for memory and search features and build evaluation systems that maintain quality as the platform evolves

• Enhance search performance across conversations using embeddings, hybrid retrieval, ranking strategies, and user-facing presentation

• Collaborate across product, design, privacy, legal and safeguards teams to ship features that handle user consent, enterprise data controls and sensitive information responsibly

• Set quality standards for the team through design feedback, code review and expectations around performance, polish and reliability

• Participate in on-call rotation for systems you own

What you'll bring

• Hands-on experience building and operating products used by large numbers of people, including the backend infrastructure supporting them

• Experience developing agentic systems or LLM-powered products, with familiarity in tool use, retrieval mechanisms and prompt design

• Strong programming ability in Python and TypeScript, with proven work on both backend services and web frontends

• Demonstrated ability to take a product from concept through launch and improve it based on usage patterns and user feedback

• Experience defining metrics and running experiments to inform product decisions

• Sharp product judgment: you spot when something doesn't work right and you fix it; you can articulate why design decisions matter

• Clear communication skills and a track record of collaborating closely with product managers, designers, and other non-engineering partners

• Genuine interest in AI safety and Anthropic's mission

Nice to have

• Founding experience, early-stage startup work, or a track record of launching products from scratch

• Background in search or retrieval systems such as embeddings, vector databases, hybrid retrieval or ranking algorithms

• Experience building evaluation frameworks for language models or measuring retrieval and ranking quality

• Work on consumer products known for strong user retention or personalization systems

• Experience shipping features that manage sensitive data or meet enterprise requirements like encryption, data residency or compliance standards

• Background in performance-critical or low-latency product development

• Experience mentoring engineers or setting technical direction for teams

What they offer

• Annual salary range of 320,000 to 485,000 USD

• Remote work with expected presence in one of Anthropic's offices at least 25 percent of the time

• Bachelor's degree required or equivalent demonstrated experience in a relevant field