Browsing Tag

Agents

3 posts
The Future of Data Streaming with Apache Flink for Agentic AI Supporting A2A and MCP
Read More
  • 3.7K views
  • 8 minute read

The Future of Data Streaming with Apache Flink for Agentic AI

Agentic AI is moving into production. Autonomous, tool-using, goal-driven systems that need real-time data and context. Apache Kafka and Flink provide the event-driven foundation to run these agents at scale. With the new Flink Agents project (FLIP-531), Flink will natively support long-running, system-triggered AI agents integrated with LLMs, tools, and emerging protocols like MCP and A2A. This marks a major step toward reliable, enterprise-grade Agentic AI.
Read More
0
0
0
Agentic AI with Apache Kafka as Event Broker Combined with MCP and A2A Protocol
Read More
  • 18.0K views
  • 7 minute read

Agentic AI with the Agent2Agent Protocol (A2A) and MCP using Apache Kafka as Event Broker

Agentic AI is emerging as a powerful pattern for building autonomous, intelligent, and collaborative systems. To move beyond isolated models and task-based automation, enterprises need a scalable integration architecture that supports real-time interaction, coordination, and decision-making across agents and services. This blog explores how the combination of Apache Kafka, Model Context Protocol (MCP), and Google’s Agent2Agent (A2A) protocol forms the foundation for Agentic AI in production. By replacing point-to-point APIs with event-driven communication as the integration layer, enterprises can achieve decoupling, flexibility, and observability—unlocking the full potential of AI agents in modern enterprise environments.
Read More
0
0
1
Event-Driven Agentic AI with Data Streaming using Apache Kafka and Flink
Read More
  • 19.6K views
  • 12 minute read

How Apache Kafka and Flink Power Event-Driven Agentic AI in Real Time

Agentic AI marks a major evolution in artificial intelligence—shifting from passive analytics to autonomous, goal-driven systems capable of planning and executing complex tasks in real time. To function effectively, these intelligent agents require immediate access to consistent, trustworthy data. Traditional batch processing architectures fall short of this need, introducing delays, data staleness, and rigid workflows. This blog post explores why event-driven architecture (EDA)—powered by Apache Kafka and Apache Flink—is essential for building scalable, reliable, and adaptive AI systems. It introduces key concepts such as Model Context Protocol (MCP) and Google’s Agent-to-Agent (A2A) protocol, which are redefining interoperability and context management in multi-agent environments. Real-world use cases from finance, healthcare, manufacturing, and more illustrate how Kafka and Flink provide the real-time backbone needed for production-grade Agentic AI. The post also highlights why popular frameworks like LangChain and LlamaIndex must be complemented by robust streaming infrastructure to support stateful, event-driven AI at scale.
Read More
0
0
0