Model Mondays - AI Red Teaming Agent скачать в хорошем качестве

Model Mondays - AI Red Teaming Agent

Assessing risks to your models and agentic AI apps is critical to building trustworthy AI. Join us as we talk to Minsoo Thigpen about the AI Red Teaming Agent on Microsoft Foundry Learn More: https://aka.ms/MM/dec8/1 Try the Workshop: https://aka.ms/MM/dec8/w 🔎 Explore the repo: https://aka.ms/model-mondays 📢 Continue the conversation on the Discord: https://aka.ms/model-mondays/discord 📌 This session is a part of a series! Learn more here: https://aka.ms/model-mondays/rsvp Chapters: 00:05 – Welcome & Housekeeping: Code of Conduct and Session Overview 01:00 – Meet the Hosts: Nitin and Chararma Kick Off the Episode 01:32 – Today’s Topic: AI Red Teaming Agents and Guest Introduction 01:44 – Weekly Highlights: Foundry Model Updates and Ignite Announcements 03:02 – Spotlight: MR Large 3 and GPT-5.1 Codex Max in Microsoft Foundry 04:42 – Deep Dive: Azure Language MCP Integration and Privacy-First Features 06:16 – Foundry MCP Server Public Preview and Modular AI Architecture 06:49 – Foundry IQ Agent Framework: Reasoning and Retrieval Innovations 08:03 – Upcoming AI Summit in NYC and Foundry Demos 09:02 – Expert Segment: Minsu Tikpen on AI Red Teaming Agent 11:08 – What is AI Red Teaming? Concepts and Importance 16:19 – Introducing the AI Red Teaming Agent: Pirate + Foundry Evaluations 22:02 – Demo: Running Red Teaming in Microsoft Foundry 27:07 – Configuring Risk Categories and Attack Strategies 29:02 – Reviewing Results: Attack Outcomes and Mitigation Strategies 33:45 – Q&A Highlights: Integration, Labs, and Continuous Red Teaming 41:02 – Foundry Friday AMA Announcement 41:24 – Customer Story: ADA’s Scale and Multi-Agent Architecture 49:11 – Performance, Cost Optimization, and Stability with Azure AI Foundry 50:01 – Models for Beginners Curriculum Preview 57:01 – Understanding Language Models: Fundamentals and Transformer Architecture 1:06:12 – Hands-On Options: GitHub Models vs Microsoft Foundry 1:11:09 – Sneak Peek: Next Episode on Multimodal Models 1:13:05 – Wrap-Up: Feedback and Future Sessions #MicrosoftReactor #learnconnectbuild [eventID:26577]