How AI Is Reshaping Software Development

Generative AI is now a real actor in software development: it writes code, tests, documentation, even proposes architectures. Yet beyond this acceleration, one question remains: what is still essential in the act of programming? Through a journey into computing’s “Hall of Fame,” we’ll revisit the foundational values of software and see how they resonate with the new possibilities opened by AI.

From probabilistic predictions to operational decisions

Over years of developing scikit-learn and exchanging with both applied data scientists and academics, I have progressively refined my understanding on topics such as the fundamental sources of "noisy" data, how to evaluate and improve the design of predictive models with respect to ranking power and probabilistic calibration, how to handle predictive uncertainty and turn probabilistic forecasts into optimal decisions with respect to an application-specific utility function.

The goal of this presentation is to share what I learned with you and hopefully help you reflect on how to quantify the value of predictions and better use scikit-learn and other predictive modeling tools.

Partner talk by Quansight: AI Pipelines That Keep Researchers, DevOps, and Security Teams Happy

Can you bridge AI technicity and business impact?

Hiring technically strong data scientist that is able to talk to the business is as easy as finding a four leaves clover. Not to talk about hiring a full team of them. 
What if told you that with correct tooling, it’s possible to bridge AI technicity and business impact?

Partner talk by skfolio: a scikit-learn vertical for portfolio optimization and risk management

An overview of skfolio, an open-source library that leverages scikit-learn’s API for portfolio optimization and risk management. We will reflect on our journey within quantitative finance, the strength of scikit-learn as a foundation and how our work aligns with the continuing evolution of machine learning and AI techniques.

Ecosystems, Not Just Code. Building Sustainable OSS Communities

This talk explores what it takes to build and maintain sustainable open source communities, using scikit-learn to illustrate successes and lessons. It raises awareness of the challenges and complexities in contributor expectations, community building, volunteer burnout, governance tensions, and company involvement. Key issues are discussed to highlight what is needed for long-term sustainability in OSS ecosystems.

skrub: machine learning for dataframes — operate data and predictive models together

This talk presents skrub, a Python library bridging data sources and machine learning predictive models. We review the current pitfalls when operating predictive models in the real world. Then, we provide an overview of how skrub addresses some of those challenges..

train_test yourself with Skolar

Skolar—the training and certification program designed by the creators of scikit-learn—aims to define and validate essential skills for data scientists and machine learning practitioners in a rapidly shifting technological landscape.

This round table brings together three complementary viewpoints: Frédéric Pascal (Director DataIA), representing the training institutions; Frits Hermans (lead data scientist from ING), bringing the expectations and constraints of the industry; and two recently certified candidates, sharing their practical experience with Skolar’s learning materials and exam pathway.

By confronting the realities of education, industry needs, and practitioner experience, the session aims to outline credible, adaptable directions for future-proofing ML competencies and ensuring trust in the capabilities of tomorrow’s data professionals.