How do you know if your AI is giving you the right answer— and what happens when someone tries to manipulate it?
Join us for another exciting TechTalks powered by CrowdStrike evening where we tackle two of the most critical challenges in AI Engineering today: evaluating AI outputs and protecting your systems from adversarial attacks.
Agenda
18:00 - 18:30 - Welcome & Networking
18:30 - 19:15 - Prompt Injection: Exploiting the Trust in Large Language Models (Bogdan Nicolae, Sr. Engineer - Frontend @ CrowdStrike Charlotte AI)
Understanding and defending against one of AI's most dangerous attack vectors
As AI systems become more embedded in our workflows, they also become a growing target for adversarial attacks. This session explores prompt injection — a technique where malicious inputs manipulate AI models into ignoring their original instructions and executing unintended, potentially harmful actions.
19:15 - 20:00 - Is It Good Enough? Mastering AI Output Evaluation (Andrei Saceleanu, Cloud Engineer II @ CrowdStrike Charlotte AI)
A practical guide to measuring, validating and trusting your AI-generated content
AI can generate outputs in seconds — but how do you know if they're accurate, reliable, or useful? This session will provide a practical overview for evaluating AI-generated content, identifying common failure modes, and building confidence in the results you act on.
20:00 - 20:30 - Networking & Discussions
What's Included
- Expert insights from CrowdStrike's AI team
- Networking opportunities with fellow developers
- Refreshments provided
Event Details
Date: Thursday, June 25th
Time: 18:00 - 20:30
Venue
CrowdStrike Bucharest HQ
America House Building
Piața Victoriei, Buchar
