Women in Cloud - Building Resilient AI

We’re excited to bring our community together again — this time with a focus on Building resilient AI.

As LLM-powered applications move from experimentation to production, a new challenge emerges: how do we secure what we don’t fully understand yet?
In this session, we’ll explore two critical layers of AI defense—protection and proactive testing—through real-world practices and practical demos.

Topics
1. Shielding your LLMs with AWS Bedrock Guardrails
Ever felt like deploying an LLM is a bit like letting a genius but unpredictable intern talk to your customers? Join me as we explore how to move past the unregulated nature of unregulated AI using AWS Bedrock Guardrails. We’ll dive into the six pillars of AI safety and deploying guardrails to build AI that isn't just smart, but is fundamentally trustworthy and brand-safe.

About the speaker
Maitreyi Singh is a Data Scientist at REA Group, currently working on developing AI agents and intelligent systems. She has experience in Generative AI and enjoys exploring innovative applications of AI in real-world solutions. Outside of her professional work, Maitreyi enjoys painting as a creative hobby.

2. Red Team Your LLM App Before the Bad Guys Do

As LLM-powered apps become more common, they're also becoming bigger targets. Prompt injection, jailbreaks, and harmful content exploits are real threats — and most developers don't discover them until it's too late. In this talk, we'll cover why guardrails are essential for any production LLM app, and why red teaming is one of the most effective ways to uncover vulnerabilities before bad actors do. We'll then put theory into practice with a live demo: a vibe-coded AI app on Amazon Bedrock, attacked using an open source red teaming tool — first without guardrails, then with them. The difference will speak for itself.

About the speaker

Vanessa is a Senior Product Security Engineer at REA Group with five years of experience spanning cloud infrastructure security, security tooling and automation, and AI security. She specialised in building secure systems end-to-end, automating security workflows, wiring up tools to catch issues earlier in the pipeline, and exploring the evolving security challenges that come with AI-powered applications. She loves getting into the weeds of how things break — and helping people build them more securely.

Agenda
5:30pm – Food and networking
6:00pm – Presentation 1. Shielding your LLMs with AWS Bedrock Guardrails
6:30pm – Presentation 2. Red Team Your LLM App Before the Bad Guys Do
7:00pm – Networking

We’re incredibly grateful to our sponsor — REA group for supporting our community and making this event possible.

Join us for an evening of learning, meaningful conversations, and great company, we can’t wait to see you there!