Harnessing Event-Driven and Multi-Agent Architectures for Complex Workflows in Gen...
summary
Proposes combining event-driven streaming (Apache Kafka, Pulsar, brokers) with multi-agent orchestration (Autogen, CrewAI, Strands) and RAG backed by vector DBs to handle stateful enterprise workflows; claims this enables real-time, decentralized decision-making and scalable tool integration while agents fill memory and preference gaps. Demonstrations are experimental (live streams, prototypes) not production; key challenges include observability, debugging distributed event flows, prompt-injection and security, and the need for MCP/A2A protocols to coordinate low-level APIs and agent-to-agent messaging.
AI-written summary · about
chapters (14)
- 0:00Event-driven & multi-agent intro
- 2:07AI/ML & deep learning overview
- 6:15Limitations of large language models
- 10:30Agentic architecture core
- 12:36Event-driven for AI workflows
- 16:24Event sourcing & CQRS in AI
- 19:10Streaming GenAI prompts
- 23:22Retrieval augmented generation
- 27:40Multi-agent systems fundamentals
- 29:23MCP & A2A protocols
- 31:31Combining event & agent systems
- 33:59Agent design patterns
- 38:37AI agent security & ethics
- 40:22Q&A
related talks
32mEvent-Driven & Multi-Agent Architectures for Gen AI Workflows
51mOptimizing Complex Workflows with Event-Driven Multi-Agentic Approach
59mOptimizing Complex Workflows with Event - Driven - Multi Agentic Approach
56mDesign Patterns for Multi Agent Systems
27mAI-Powered ETL Pipeline Orchestration: Multi-Agent Systems in the Era of Generative AI
29mAgentic AI Meets Event-Driven Architecture
26mFrom Chaos to Choreography: Multi-Agent Orchestration Patterns That Actually Work
14mOrchestrating Agentic Workflows with Governance and Trust