Abstract
An hour with James Carter (head of CORTEX, Prosigns AI/ML practice) walking through the reference RAG architecture documented in our 38-page whitepaper. We'll cover the architectural decisions teams make at each layer: retrieval, reranking, evaluation, governance, cost, and where production RAG most commonly fails. Live Q&A in the final 10 minutes.
Recording sent to all registrants regardless of attendance. No follow-up sales sequence. We send the recording and the architecture diagrams in PDF, and that's it.
Agenda
- 01
Reference architecture: components and contracts (15 min)
- 02
Hybrid retrieval and reranking under real load (15 min)
- 03
Eval harness: faithfulness, groundedness, refusal correctness (10 min)
- 04
Cost optimization: model tiering, semantic caching (10 min)
- 05
Q&A with the head of CORTEX (10 min)