Room “U3-05” - BIOS Building (U3)
From Theory to Practice: How Algorithms and Data Structures Power Real Systems
Speaker: Dr. Marco Previtali, Bloomberg L.P.
Abstract
This talk shows why algorithms and data structures are not just abstract concepts, but essential tools for building and understanding real systems. Even if modern AI can generate code and pass tests, it still struggles to reason about scale, failures, and resource limits—areas where fundamental knowledge remains crucial.
Through a series of practical examples, I will show how simple ideas presented in the Algorithms and Data Structures course appear in real situations. For instance, finding the exact moment when a bug was introduced in a complex system can be done efficiently using an approach similar to binary search (as in git bisect), instead of checking every change. In another case, I will briefly discuss how modern search systems rely on specialized data structures for fast retrieval, and how these data structures can be seen as extensions of the basic data structures presented in the course. I also want to stress that these data structures are at the core of many AI systems today.
More generally, I will highlight how common techniques help us design systems, explore complex environments, and make better decisions under constraints. The goal is to provide an intuitive and motivating perspective on how foundational concepts translate into practice, even in the LLM-assisted coding era.
The talk is aimed at first-year Computer Science students, but it is open to anyone who is interested.
Short bio
Marco Previtali is a Senior Software Engineer at Bloomberg L.P., based in the Lugano office. He works on backend systems for PORT, Bloomberg’s Portfolio & Risk Analytics solution used for portfolio analysis, performance attribution, risk forecasting, scenario analysis, and optimization. Before joining Bloomberg, he was a Software Engineer at Meteomatics, contributing to software systems for weather and climate data products. He holds a PhD in Computer Science from the University of Milano-Bicocca, where his research focused on algorithms and data structures for bioinformatics.
contact person for this seminar: yuri.pirola@unimib.it