Post by SIDHARTH K

Attended the Agentic Football World Cup Workshop at AWS Bengaluru recently, and it turned out to be a pretty interesting way to get hands-on with agentic AI. We built a team of AI agents using Amazon Bedrock AgentCore and Strands Agents SDK, all running through Kiro. Instead of just calling APIs, the agents had to make decisions, coordinate with each other, and adapt in real time during simulated football matches — which made the learning much more concrete. What I found useful was seeing how multi-agent systems actually behave when you put them in a dynamic environment. Things don’t always go as expected, and that’s where most of the learning happens — especially around orchestration, state, and interaction between agents. Overall, a solid hands-on experience that gave a clearer picture of how agentic workflows can be built and where they might actually fit in real-world use cases.