Senior Agentic AI Engineer

The role

ConstraintFlow is looking for a Senior Agentic AI Engineer to help build the intelligence and agent architecture behind our Operational Intelligence Platform.

This role goes well beyond building chatbots or wrapping an LLM API. You will work on production agentic systems that retrieve governed context, reason across enterprise information, invoke tools, participate in workflows, interact with specialized agents and preserve the relationship between recommendations, human decisions, actions and outcomes.

You will work across Gurita, ConstraintFlow's Agent & Workflow Plane, and the broader platform architecture supporting Manufacturing and Life Sciences.

What you'll do

• Design and build production-grade AI agents and multi-agent workflows.
• Build tool-use and function-calling architectures.
• Implement retrieval and context orchestration across structured and unstructured enterprise information.
• Build short-term, episodic, semantic, procedural and temporal memory capabilities where appropriate.
• Design human-in-the-loop approval and escalation patterns.
• Implement agent state, workflow state and durable execution.
• Build evaluation frameworks for agent quality, reliability and safety.
• Develop observability for reasoning workflows, tool execution and failures.
• Implement authorization boundaries for agents, tools and knowledge.
• Work with Zahir, ConstraintFlow's governed Data & Context Plane.
• Integrate agents with APIs, enterprise applications and event-driven systems.
• Improve latency, reliability and cost of production agent workflows.
• Contribute reusable agent capabilities rather than one-off customer bots.
• Work with product and domain teams to translate real operational workflows into agent behavior.

Technical profile

Strong experience with:

• Python
• LLM application architecture
• Agentic AI systems
• Tool / function calling
• RAG, vector retrieval and structured retrieval
• APIs and event-driven architectures
• Distributed systems
• Production AI evaluation
• Cloud deployment

Useful experience

We care more about evidence that you have built reliable AI systems than whether you have used every framework in our stack.