Senior Agentic AI Engineer

This role is for engineers who think in architecture, algorithms, and systems, not prompt chains. We expect you to understand how and why agentic AI works, not simply assemble graphs. You should be comfortable designing new execution models, planning strategies, memory architectures, and coordination mechanisms when existing abstractions become limiting.

Responsibilities:

• Design autonomous planning and reasoning systems.
• Build cognitive architectures for enterprise decision-making.
• Design working, episodic, semantic, and organizational memory.
• Build distributed multi-agent systems.
• Design autonomous execution and recovery mechanisms.
• Build evaluation systems that continuously improve agent behavior.
• Translate research into production-grade systems.
• Proprietary reasoning and planning architecture.
• Long-horizon autonomous agents.
• Enterprise memory systems beyond document retrieval.
• Multi-agent coordination and negotiation.
• Self-evaluating and self-correcting agents.
• Reliable autonomous execution in production.
• Core IP that defines the future of enterprise AI.

This role is probably not for you if:

• Your primary AI experience is building RAG applications.
• You primarily build chatbot or workflow wrappers around foundation models.
• Your architecture starts and ends with a vector search.
• Your approach to "agents" is chaining prompts together.
• Your primary engineering environment is low-code agent builders or visual workflow tools.

Requirements:

• Required Expertise: Planning and Search Algorithms, Multi-Agent Systems, Distributed Systems, State Machines, Knowledge Representation, Optimization, Decision Theory, Reasoning Under Uncertainty, Asynchronous Systems, and AI Evaluations.
• Absolute expertise with LangGraph is expected; dependence on any single framework is not.

You Should Be Comfortable Discussing:

• Planning versus reactive execution.
• Centralized versus decentralized coordination.
• Memory retrieval versus memory synthesis.
• Symbolic versus LLM reasoning.
• Long-horizon execution.
• Agent reliability and failure recovery.
• Enterprise safety and governance.