Hello, I’m Aicha Zeroual — a Big Data and Cloud Computing Engineering student at ENSET Mohammedia in Morocco. I’m passionate about data, AI, and software development, and I’m always curious about how technology can be used to solve real-world challenges.
Beyond coding, I’m deeply interested in collaboration, knowledge-sharing, and public speaking. I love connecting with others in tech, learning through meaningful conversations, and contributing to projects that make an impact.
AI systems don’t fail like traditional software. They often continue making predictions long after the world they were trained on has changed — silently degrading in performance. In this talk, we’ll explore how and why models break after deployment, and what it means when they no longer reflect reality.
This session is perfect for anyone interested in understanding how to keep AI models reliable after production.
What you will learn:
- Why monitoring drift is critical to building trustworthy AI systems
- The difference between data drift, training-serving skew, and concept drift.
- Real-world examples of drift and its impact on model performance
- How to detect drift with and without labeled data
Join me to understand how to recognize and respond when your model’s intelligence starts to freeze.
Searching for speaker images...