Senior AI Engineer - Sana

You'll build the core agent infrastructure that powers Sana's mission to bring superintelligence to work. This is a greenfield opportunity to define how AI agents plan, reason, and execute across enterprise environments—building systems that reliably handle real-world complexity at scale.

Responsibilities

Architect multi-step planning, orchestration, and tool routing for agents; Implement code generation agents and sandboxed code execution; Engineer memory, state, and context packing/grounding strategies; Balance latency, quality, and cost controls for agent execution; Develop safe fallbacks, graceful degradation and robust error handling; Collaborate with platform and search teams to deliver reusable agent infrastructure; Establish safety guarantees and measurable quality improvements

Requirements

3+ years of software engineering experience building production backend or platform systems; 3+ years of experience in TypeScript, with a strong track record of writing reliable, maintainable services; 3+ years of experience with distributed systems, APIs, asynchronous workflows, and service-oriented architecture; 3+ years of experience designing systems with a focus on scalability, reliability, observability, and maintainability

Nice to have

Experience building and deploying LLM-powered applications in production; Experience building agent platforms or AI infrastructure; Deep understanding of the low-level details of the OpenAI, Google, and Anthropic LLM APIs, including tool calling, system prompt caching, etc.; Familiarity with LLM application patterns, including tool calling, retrieval-augmented generation (RAG), memory and context management, multi-step orchestration, and human-in-the-loop systems; Experience building and running machine learning systems in production, including compiling training and test datasets, building training pipelines, evaluating models, and detecting and handling drift (neural networks, Gaussian models, Thompson sampling, etc.); Experience designing evaluation frameworks for LLM or agent quality and safety, including hands-on use of platforms such as Langfuse or LangSmith; Familiarity with vector databases, prompt and context engineering, and experimentation tooling; Experience working with sandbox environments such as Modal, and designing strict access control models to keep user data safe and encrypted at all times; Experience running services in Kubernetes-based environments on GCP or equivalent cloud platforms; Comfort working with Postgres and Redis in high-throughput, low-latency service contexts; Contributions to open source TypeScript projects; Ability to navigate ambiguity, make strong technical tradeoffs, and drive projects from concept to production; Strong communication and collaboration skills, with the ability to partner effectively across engineering, product, and AI research teams

Benefits

Eligibility for Workday Bonus Plan or role-specific commission/bonus; annual refresh stock grants; comprehensive benefits; Flex Work (at least 50% of time in office or with customers/prospects/partners)