AI Agent Orchestrator (Python or Go)
Company: A sovereign wealth fund-backed Saudi Arabian state AI project.
Location: Fully remote (Europe / CIS). Relocation to KSA is optional and discussed separately.
Role Summary: We need an engineer who uses agents as their primary work tool daily. Someone who can take an ambiguous task and transform it into a working multi-agent system: assign roles, set up orchestration, deploy to production, and ensure it operates reliably.
Responsibilities:
- Design and build multi-agent systems in production: tool calling, memory, structured outputs, human-in-the-loop at necessary points.
- Set up orchestration using LangGraph / MCP / Claude Agent SDK or similar tools.
- Configure evaluation and observability: LangSmith, LLM-as-judge, chain tracing, quality control.
- Integrate agent systems with real business processes — APIs, queues, databases, access rights.
- Define the boundary between deterministic logic and LLM, build reliable fallback scenarios.
Requirements:
- ! Strong backend foundation: 8+ years in Go or Python, queues, production systems !
- Daily use of agents: Claude Code, Codex, Cursor, or a custom pipeline as the primary workflow.
- Practical experience with multi-agent systems in production, real systems under load.
- Understanding of LLM limitations and the ability to build protective measures around them.
- English B2+
- Willingness to enter into a B2B contract (possible from RF or RB).
Conditions:
- Long-term B2B contract
- Salary range: from €10,000/month
- The team is being formed from scratch, 10 people, high ownership.
- Remote work, relocation to KSA is optional.
