Code with Cosmos: Operating AI Agents in Production @MicrosoftDeveloper
Code with Cosmos: Operating AI Agents in Production  @MicrosoftDeveloper
Uploaded October 2026 | Updated November 2026, 3 hours ago
How do you take an AI agent from a working demo to a production-ready system? Marko Hotti and Madeline Egan walk through a practical framework to observe, evaluate, and optimize AI agents with Azure Cosmos DB. Explore the full repo: https://aka.ms/cosmos-travel-planner Watch the full series: https://aka.ms/CodeWithCosmosPlaylist

Learn how traces, evaluation suites, regression testing, and before-and-after metrics can help improve retrieval, latency, token usage, memory efficiency, and Azure Cosmos DB request unit consumption across the data, retrieval, memory, and LLM layers.

đź’» GitHub repo: https://aka.ms/cosmos-travel-planner
▶️ Watch the full series: https://aka.ms/CodeWithCosmosPlaylist
Try Azure Cosmos DB for free: https://aka.ms/trycosmosfree
Code with Cosmos: Operating AI Agents in ProductionQDK Tutorial: Build Chemistry applications with the Quantum Development Kit for ChemistryMake Machines Smarter with Physical AIMicrosoft Foundry: Test Your Agent Before You ShipWhat’s it like for Microsoft to manage open source projects?  What’s in it for Microsoft?Fast Focus: What’s New Across Microsoft SQL and Developer Tooling in 2026Stop Writing Prompts. Start Writing Specs.TypeScript 7 Is Here (And Its 10× Faster).NET Memory Dumps, Database Time Migration & AI Agent MemoryThe Developers Guide to Azure SQL Hyperscale4X the Turns, 16X the CostZoomIt for Mac
Microsoft Developer |

Code with Cosmos: Operating AI Agents in Production

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER