Building Intelligent AI Infrastructure with O.XYZ's ORI
summary
ORI routes decomposed user queries as a chain of atomic queries, sends each subquery to the top-performing open‑source LM (matched by a benchmark matching engine tied to Hugging Face leaderboards), and runs on H100 clusters via Ray Serve and vLLM with i.net for distributed compute. It fine‑tunes the most‑routed models, enforces ethical filters and compliance metadata (SOC2/HIPAA style), demos routing code to CodeLlama, claims 10–20× faster inference, and reports outperforming a single 70B model on multiple benchmarks.
AI-written summary · about
chapters (10)
- 0:00Introduction to O.XYZ
- 1:02Industry challenges in AI
- 2:35Core features of ORI
- 5:54Benchmark and routing engine
- 11:11Agentic workflows
- 12:23Technical infrastructure
- 13:43System insights and vision
- 15:07Performance results
- 15:55System demonstration
- 19:14Partnerships and closing
related talks
28mNavigating LLM deployment: Tips, tricks, and techniques
27mAI Demystified Unveiling the Intelligence Behind the Machine
01h 01mBuilding the Intelligent Enterprise: Ensuring Success with Data and GenAI
21mAI-Powered O11y with LLMs: Resilience in Cloud PaaS Archs
29mOpen Source Sovereign AI: Building Enterprise GenAI Platforms for Control & Comp...
35mIntroducing Open Platform for Enterprise AI
25mFrontier AI Models: A Review of Past Nine Months
39mBuild your enterprise Generative AI applications with Microsoft Olive