Fullstack Software Engineer, ADK

What you'll do

• Develop end-to-end agent features across tool/MCP and orchestration backends and the React console.
• Build runtime capabilities including durable execution, retries, scheduling, and human approvals; improve agent evaluation to catch behavioral regressions.
• Create a real-time console for agent events, tool calls, and approvals, and establish reusable engineering patterns and specifications.

What they ask for

• Strong software engineering ability and experience building real full-stack products with a Python backend and React frontend, or ability to reach that level quickly.
• Rigor in specifying and evaluating non-deterministic systems, and ownership of production outcomes, including AI-assisted code.
• Understanding of how agent frameworks, tools, and evaluations work, and regular use of AI coding tools.
• Ability to work effectively amid fast-moving ambiguity.

Nice to have

• AI/agent product experience, including MCP, agent frameworks, or evaluations; durable workflow engines such as Temporal; streaming UIs; or OpenTelemetry.
• Spec-driven development, multi-tenant SaaS, or aerospace, defense, or discrete-manufacturing context.

Tools

Python, React, MCP, Temporal, OpenTelemetry

AI focus

Build agent capabilities and orchestration, harden agent execution and approvals, and develop evaluation systems to assess agent behavior before release.