Member of Technical Staff, Agent Harness
Hark, San Jose +1
The role
You'll build the backend systems that make Hark's AI agent actually work — reliable, fast, and production-grade.
That means the hard infrastructure problems: high-concurrency services, low-latency streaming, state management for long-running agent workflows, and the execution layer that connects model outputs to real-world actions. You are building the nervous system of an agentic product.
This is a high-ownership role on a small team. You'll work directly with model researchers and platform engineers, and the systems you build will determine whether Hark feels like a slow chatbot or something genuinely new.
What you will do
- Core Runtime Architecture: Design and build the high-concurrency backend services responsible for agent orchestration, tool execution, and real-time streaming.
- Systems Reliability: Develop robust failure-handling, retry logic, and state management for long-running agentic workflows.