Skip to content

Details

Agenda:

Session 1: From Log Hunting to AI-Powered Insights: Building Event-Driven Observability on AWS Serverless by Sourabh Choubey, Lead Application Developer at Schneider

Abstract:
Most teams know when something breaks. Very few know why — quickly.
This session walks through a production-grade, event-driven Root Cause Analysis pipeline built entirely on AWS serverless — no third-party APM tools, just CloudWatch, EventBridge, Step Functions, Lambda, and Bedrock working together.

Part 1 — The Foundation: Structured error classification with EMF metrics, typed CloudWatch alarms, and a Composite Alarm that auto-triggers a Step Functions investigation — logs delivered in 90 seconds.

Part 2 — The Brain: Amazon Bedrock plugged into the same pipeline. Two emails land simultaneously — raw logs in ~10 seconds, AI-generated RCA with root cause, impact, and fix recommendations in ~30 seconds.
Real failure scenarios. Real production trade-offs. No guesswork.

Session 2: AI Agents with Amazon Bedrock Agentcore by Pratik Raj, Senior Developer at SLS Guru

Abstract:

  • The talk explores the shift from chatbots to action-oriented, agentic AI
  • Agents are built on an architecture of LLMs, tools, and a Reasoning → Action loop
  • Enterprise adoption requires addressing security, establishing guardrails, and human oversight
  • Amazon Bedrock AgentCore offers secure, serverless environments for persistent agent execution
  • It also discusses Agentcore for web automation and the Strands open-source SDK

Related topics

Artificial Intelligence
Amazon Web Services
Cloud Computing
Serverless Architecture
Software Development

You may also like