74th Data Science Leuven meetup

19h: Doors

19h30:

• Shubhangi Goyal, Context driven AI Agents, In this session, I will explore how agents are built and use dynamic context to execute complex tasks and deliver real value. This also includes using context engineering to move beyond single prompt based interactions.
• Dionys Nabaro, Why Forecasting Isn’t Enough: Modelling Procurement Transparency at Predikt, Forecasting commodity prices is harder than it looks — as we found out first-hand during one of our internal hackathons. Rather than chasing increasingly complex models for marginal gains, we’ve learned to focus on something more actionable: transparency. In this talk, I’ll share how we think about modelling procurement data at Predikt, and why giving teams a clear, connected view of their costs and market signals often matters more than the forecast itself.
• Manuel Alejandro Ledezma Falcon, Machine Learning for Testers: Train, Evaluate, and Detect Bias in AI Models, This session introduces software testers to the fundamentals of Machine Learning from a QA perspective through hands-on Python examples such as training and evaluating a simple spam classifier on real-world data. It focuses on how to test ML systems in practice, including data requirements, model validation, and the identification and mitigation of bias to ensure fair, robust, and trustworthy AI.

The last session will include a live workshop, so you're invited to bring your own laptop to test things out. We've reserved 15 minutes extra to accomodate for this.

21h15: Beers (STUKcafé, Naamsestraat 96, 3000 Leuven)

We're always on the lookout for interesting presentations for future sessions. Submit your talk here: https://sessionize.com/data-science-leuven