Software Engineer, AI Platform

Vanta is hiring a Software Engineer, AI Platform to build shared foundations that help teams deliver reliable AI products. The role focuses on AI platform infrastructure, agent workflows, evaluation and debugging tools, and operational ownership in partnership with product teams.

Responsibilities:

• Build shared SDKs, APIs, and infrastructure for AI features and agentic workflows
• Integrate durable execution infrastructure for long-running agent workflows, including checkpointing, retries, human-in-the-loop steps, and recovery from partial failures
• Evolve our agent harness, the runtime for tool calling, context management, memory, and model orchestration
• Develop evaluation and debugging tools that turn production failures into measurable improvements and regression coverage
• Improve model integrations, retrieval, and context delivery to support accurate, useful AI behavior
• Own platform capabilities through rollout and operation, including reliability, observability, performance, and secure handling of customer data
• Partner with product teams to prototype solutions, identify common needs, and drive adoption of reusable platform capabilities

Requirements:

• Strong software engineering fundamentals and experience building and operating backend services, APIs, or developer platforms
• Hands-on experience shipping LLM-powered products or agentic workflows and understanding their failure modes
• Experience with durable execution or workflow orchestration platforms and designing for idempotency, retries, and long-running state
• Familiarity with agent harnesses and frameworks, including tool design, context management, subagents, and their tradeoffs
• Practical evaluation experience: defining success, diagnosing failures, and measuring whether changes improve outcomes
• Product judgment: connecting platform decisions to developer needs and customer outcomes
• Ownership, curiosity, and clear communication across teams in a rapidly changing technical environment
• Uses AI in day-to-day work to amplify their skills and strengthen their outcomes, applying it across routine workflows, actively building fluency through experimentation, and exercising sound judgment to ensure AI-assisted work meets a high quality bar