Sudhir Reddy Maddulapally

Agentic AI systems autonomously orchestrate tools, integrate third-party components, and make decisions, introducing risks from compromised dependencies and agents deviating from intended behavior. Protecting prompts, agent memory, retrieval context, and model weights requires strong isolation and verifiable trust.

Highly sensitive workloads often rely on cryptographic proof about the code, config and other aspects of the execution environment before unsealing code, data to be processed.

There are numerous implications to data privacy and intellectual property security in generative AI applications, here we will focus on solving for some of those using trusted executions environments (TEEs).

In this workshop, learn the various ways that AWS Nitro Enclaves can help improve the security of LLM-based applications by protecting sensitive data and models while they are being used. Having a background in application development, systems administration, and information security is not required but is recommended.

Data protection is crucial when it comes to processing sensitive datasets for healthcare, financial, demographic, and social media use cases. In this workshop, learn to build and operationalize your AWS Nitro Enclaves applications using CI/CD.