Post Content
Optimize enterprise Agentic AI with a tiered system-of-models architecture in Microsoft Foundry. This session demonstrates a plan-and-execute pattern routing tasks across frontier models for reasoning, NVIDIA Nemotron for complex sub-tasks, and local models for latency-sensitive execution. Learn to route workloads across cloud and edge tiers to minimize cost-per-task while maximizing quality. Dive into using special agents and post-trained open-source models to achieve faster task completion.
𝗦𝗽𝗲𝗮𝗸𝗲𝗿𝘀:
* Aysen Ilkbahar
* Stephen McCullough
* Joey Conway
𝗦𝗲𝘀𝘀𝗶𝗼𝗻 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻:
This is one of many sessions from the Microsoft Build 2026 event. View even more sessions on-demand and learn about Microsoft Build at https://build.microsoft.com
BRKSP94 | English (US) | Agents & apps
Breakout | (300) Advanced
#MSBuild
Chapters:
0:00 – Introduction to NVIDIA announcements and partnerships
00:05:46 – Overview of Nemotron open model family and capabilities
00:07:52 – Announcement of Nemotron 3 Ultra – NVIDIA’s most capable open model
00:08:55 – Emphasis on Open Publication and Community Confidence
00:18:35 – Vision of digital workforce and agent-based future of companies
00:18:54 – Transition to Hermes orchestration overview and handover to colleague
00:26:31 – PR creation completed and demonstration of observability and audit trail for enterprise agents
00:27:15 – Reviewing PR and Monitoring Changes
00:34:35 – Benefits of Foundry Hosted Hermes Agents and Persistent Learning Read More Microsoft Developer
