Principal Architect - Gen AI

Responsibilities:

Design and build the master orchestrator and its specialised sub-skills: routing an input to the right patent-legal skill, defining skill contracts, and composing partial outputs into one coherent deliverable.

Build RAG pipelines grounded in the CDL and a patent/prosecution knowledge graph, so every generated assertion traces to a located source rather than being produced free-hand.

Make grounding and hallucination control structural: back-checking generated claims and arguments against source disclosures, and enforcing attorney-grade QC.

Own the evaluation strategy for output you cannot fully trust: measurement without abundant ground truth, abstention and escalation paths, and confidence thresholds that decide when to route to a human.

Design human-in-the-loop control: an autonomous flow that pauses at the right attorney checkpoints and remains fully steerable.

Work with model selection, prompt architecture, structured output, and, where warranted, fine-tuning on the patent corpus.

Requirements:

Hands-on production experience building agentic or multi-agent LLM systems: orchestration, tool-calling, RAG, and retrieval, not only prompt engineering.

Fluency with the current toolkit: frameworks such as LangGraph, CrewAI, or equivalents; vector stores; and observability/evaluation tooling.

A real grasp of evaluation for generative systems, including how to measure quality when labels are scarce and errors are asymmetric.

Python depth, and comfort owning a system from design through production.

Bonus: knowledge graphs or GraphRAG; fine-tuning; any exposure to legal, compliance, or other high-stakes document domains.