Lead AI Architect - Agentic Systems & LLM Orchestration 

The Opportunity 

We're seeking a Lead AI Architect to design and build the orchestration of ATOM. You'll architect for multi-agent systems that decompose complex engineering problems, coordinate specialized tools and models, and implement sophisticated feedback loops that continuously improve first-pass success rates. 

What You'll Own 

• Multi-agent system design using LangGraph, AutoGen, CrewAI, or equivalent, with explicit state machines, retry logic, and checkpoints 

• Master orchestrator architecture that coordinates specialized agents, manages dependencies, and defines clear agent contracts and tool interfaces 

• Structured decomposition pipelines that parse ambiguous natural-language inputs into actionable engineering constraints 

• Closed-loop evaluation and error correction systems combining rule-based fixes with LLM-driven patching, confidence scoring, and escalation 

• Latency and cost optimization across the entire orchestration pipeline 

• Production monitoring, logging, failure analytics, and continual learning from real-world deployment data 

Required Experience 

• Proven track record building production agentic AI systems (LangGraph, AutoGen, CrewAI, or custom implementations) 

• Deep expertise in LLM behavior, RAG systems, prompt engineering, and multi-step tool orchestration 

• Hands-on experience with LLM fine-tuning and model evaluation 

• Strong Python systems engineering background with experience designing fault-tolerant, multi-step workflows

Strong Plus 

• Experience building code-generation or constrained generation systems 

• Background in robotics, simulation, motion planning, or compiler design 

• Experience optimizing inference latency and cost at scale