Build agents that use tools, follow plans and stop where the stakes demand it.
About the Role
Build production AI agents: systems that decide their next step and act inside real business systems, with the engineering that makes that trustworthy.
What You'll Do
Design agent architectures: planning, memory, tool use
Build tool layers with strict contracts and error semantics
Implement layered guardrails including injection defence
Build golden-set evaluation and run it on every change
Design shadow-mode rollouts before autonomous execution
Instrument audit logging for every action taken
What You'll Bring
2+ years software engineering, some of it on AI systems
Strong Python or TypeScript
Hands-on with agent frameworks and function calling
Understanding of prompt injection and mitigation
Scepticism: you assume the model will do something stupid
Nice to Have
LangGraph or Model Context Protocol experience
Evaluation tooling experience
Production experience with agents that take real actions
What You'd Build
These aren't hypothetical projects, they're live products you can try before your first interview.
