Orchestrating Intelligent Agents
summary
Over two years a working multi-agent system is built that treats agents as both clients and servers exposing tool definitions, orchestrates complex tasks into runs and threads, and uses human-in-the-loop where needed. Azure Cosmos DB is the persistence layer to persist run and thread context and store agent/tool definitions, chosen for flexible schema, low-latency global distribution and real-time consistency. The claim is that this setup makes multi-agent collaboration scalable, instantly consistent across agents and easy to experiment with as protocols evolve.
AI-written summary · about
related talks
56mBuilding AI Agents and Multi-Agent Systems with Azure AI
26mAI powered automation & multi-agent orchestration
28mThe Put That Sh*t On Agent: Building Multi-Agent AI Systems That Actually Talk to Each Other*
46mAI Agents for the Rest of Us
58mAdvanced Autonomous AI Agents: Memory, Collaboration and Evolution in Action
52mDesigning Production-Ready Agentic Systems: Lessons from the Field
45mAn Agent Walks Into a Bar... and Meets Another Agent: Multi-Agent Systems in Action
39mEvolving Intelligence: Designing Autonomous AI Agents