Software Engineer, AI Platform

Mixpanel Mixpanel · Data AI · San Francisco, CA · ENG / EPD - Engineering

Software Engineer for a newly formed AI Platform team focused on building the infrastructure to accelerate AI agent development at Mixpanel. This includes designing backend services, agent orchestration frameworks, evaluation systems, and serving infrastructure.

What you'd actually do

  1. Platform Architecture: Design and develop the core backend services, APIs, and microservices that enable product teams to easily and securely leverage AI models.
  2. Agent Orchestration: Build scalable frameworks and tools to support multi-step agent workflows, including task decomposition, tool invocation, and persistent memory.
  3. Evaluation & Reliability: Build robust evaluation systems to continuously measure reasoning quality, hallucination rates, and task success.
  4. Operational Excellence: Architect high-performance serving infrastructure with strict guarantees around latency, throughput, cost-efficiency, and error handling.
  5. Observability & MLOps: Ensure comprehensive monitoring, structured logging, and distributed tracing across all deployed AI models.

Skills

Required

  • Software engineering experience
  • Full-stack fundamentals
  • Debugging and technical investigation skills
  • Technical communication
  • Ability to move fast and iterate in ambiguous environments
  • Hands on experience integrating and orchestrating LLMs and agents
  • Experience building AI native systems, including iteratively improving and scaling them in production
  • Familiarity with techniques used to optimize AI agents: eval frameworks, agent tooling, vector search engines, context engineering, and prompt engineering.

Nice to have

  • experience collaborating in an asynchronous remote environment

What the JD emphasized

  • Hands on experience integrating and orchestrating LLMs and agents
  • Experience building AI native systems, including iteratively improving and scaling them in production
  • Familiarity with techniques used to optimize AI agents: eval frameworks, agent tooling, vector search engines, context engineering, and prompt engineering.

Other signals

  • building infrastructure for AI product development
  • accelerating AI agent development
  • building scalable frameworks and tools for multi-step agent workflows
  • building robust evaluation systems for AI agents
  • architecting high-performance serving infrastructure for AI models