ENGINEERING
Agentic Systems Engineer
Design the multi-agent systems our clients use to automate real work: agents that plan, call tools, check each other and escalate to humans. This is frontier engineering with enterprise consequences — reliability is the product.
- Remote
- Engineering
- Full-time
What you’ll do
- Design agent architectures for client processes: planning loops, tool catalogues, memory and state
- Build MCP servers and integrations that expose client systems to agents safely
- Implement observability for agent runs — traces, replay, cost and failure taxonomies
- Set autonomy boundaries with clients and encode them as enforceable policy
- Measure agent reliability against the human baseline and iterate until it clears it
What we’re looking for
- 3+ years of backend engineering, with production experience building agent or workflow systems
- Hands-on with agent frameworks and protocols — LangGraph, Claude Agent SDK, OpenAI Agents, Model Context Protocol (MCP)
- Strong TypeScript or Python, and experience with durable execution (Temporal, queues, sagas)
- Understanding of tool permissioning, sandboxing and human-approval gates
- Experience evaluating agent trajectories, not just single completions
What we offer
- Competitive salary
- Health and wellness benefits
- Remote work flexibility
- Frontier-model API budget
- Stock options