Our Event Speakers

Generative AI’s power stems from innovation across the entire tech stack — from custom silicon and cloud infrastructure to advanced multimodal models handling text, images, and audio. This session unveils how AWS leads at every layer: from Trainium and Inferentia chips, to scalable storage with S3, high-performance training on SageMaker, and model deployment via Bedrock. Through real-world examples and demos, you’ll learn how to build intelligent, multimodal AI systems and accelerate innovation from silicon to superintelligence. Whether you’re an architect, data scientist, or AI enthusiast, gain a clear roadmap to harness AWS’s AI ecosystem for next-gen solutions.

Managing secrets securely is critical for any DevOps or cloud infrastructure workflow. This presentation provides a practical guide on how to safely store secrets in Git repositories and use them in AWS via Terraform—without compromising security. We will explore best practices for encrypting and managing secrets. Attendees will learn how to encrypt secrets, deploy them to AWS Secrets Manager and SSM Parameter Store, and retrieve secret values in various AWS services.

This presentation explores how Explainable AI (XAI) can enhance AWS GuardDuty by making its threat alerts more transparent and trustworthy. The study, submitted to the Springer Nature Journal of Supercomputing, integrated SHAP and LIME into a GuardDuty workflow and tested them with security analysts. Results showed that XAI-supported alerts improved classification accuracy, analyst confidence, and reasoning quality. LIME offered clear, per-alert explanations, while SHAP provided broader, model-level insights. The talk demonstrates how applying XAI to GuardDuty can reduce false positives, boost trust, and accelerate incident response, transforming it from a simple detection tool into a reliable decision-support system for cloud security.

Learn how to create, deploy, and manage intelligent agents using the new Strands Agents SDK. In this session we'll review agents, introduce the SDK and demonstrate Python examples for agent development.

S3 buckets are one the most used storage services provided by AWS. Uploaded files are accessed as keys, or objects in AWS lingua. You can enable S3 server access logging on a bucket and start delivering logs to another destination bucket. However, the sheer volume of logs make it cost prohibitive because they come in plaintext format. Wouldn't be wonderful to be able to start analyzing access patterns to find unused objects or track operations performed by leaked credentials? In this presentation, I will go over an automated process we have developed to convert day's worth of S3 server access logs logs into parquet format reducing storage footprint by around 90%. We will delve deep into technical details in order to setup infrastructure to query data and convert objects using Athena queries.