AI agents extend the capabilities of LLMs by breaking down goals, making decisions, and handling exceptions to accomplish complex tasks.
Multi-agent systems divide complex tasks among specialized, autonomous agents that collaborate to achieve a larger goal.
The Model Context Protocol (MCP) and Agent-to-Agent (A2A) protocol provide standardized interfaces for agents to interact with resources and each other.
Agentic systems can be structured using patterns like orchestrator-worker, hierarchical agents, blackboard, and market-based models to manage collaboration.
Agent behaviors are implemented using patterns like reflection for self-critique, tool use for external data access, and planning for task decomposition.
Event-driven architecture (EDA) helps manage the dynamic flow of data in AI systems, enabling asynchronous communication for better scalability and flexibility.
The CAP theorem and PACELC framework explain the trade-offs between consistency, availability, and latency in distributed agentic systems.
Implementing enterprise-level agentic AI requires treating scalability, concurrency, and resiliency as core business concerns, not just technical challenges.
Jobs that call for the skills explored in this talk.