Software Engineer, Agentic Infrastructure

Handshake · Enterprise · San Francisco, CA · Engineering

Software Engineer focused on building the core agent orchestration layer, including tool use, memory, and multi-step reasoning systems, to power AI-driven features for millions of users. The role also involves designing evaluation, observability, and reliability frameworks for agent behavior and establishing engineering standards for agentic development.

What you'd actually do

  1. Architect & build the core agent orchestration layer — including tool use, memory, & multi-step reasoning systems — that powers AI-driven features for 20M+ users & 1M+ employers across Handshake's platform
  2. Design evaluation, observability, & reliability frameworks that ensure agent behavior is safe, auditable, & production-ready at scale
  3. Establish engineering standards for agentic development across Handshake's platform teams, enabling every engineer to build with AI
  4. Partner with ML, product, & platform engineers to ship agent-powered features from infrastructure to production at Olympic pace

Skills

Required

  • 2-4 years of backend engineering experience
  • strong proficiency in NodeJS, Typescript, &/or Python
  • Experience designing & operating distributed systems at significant scale for developer experience
  • Hands-on experience building with LLM orchestration frameworks (LangChain, LlamaIndex, or similar)
  • Deep understanding of agent architectures: tool use, multi-step reasoning, memory management, & evaluation
  • Experience with cloud infrastructure (AWS, GCP, or Azure)
  • infrastructure-as-code tools like Terraform
  • Experience with Node, JavaScript, CICD, Docker
  • A track record of shipping production systems with high reliability, low latency, & strong observability
  • Strong communication skills & ability to drive alignment across engineering & cross-functional partners

Nice to have

  • Open-source contributions to LLM or agent frameworks
  • Prior experience at an AI-native company or on a dedicated AI platform team

What the JD emphasized

  • Deep understanding of agent architectures: tool use, multi-step reasoning, memory management, & evaluation
  • A track record of shipping production systems with high reliability, low latency, & strong observability

Other signals

  • building foundational systems for AI agents
  • agent orchestration layer
  • tool use, memory, multi-step reasoning
  • evaluation, observability, reliability frameworks for agents
  • shipping agent-powered features