On June 5, 2025, Mila held the grand finale of the second edition of Speed Science, its signature science communication contest, in which AI researchers took to the stage to synthesize and communicate their research to a non-scientific audience.
Over the course of several weeks, the participants honed their skills through professional workshops and individual coaching on science communication, storytelling, and public speaking in preparation for the grand finale. Their ultimate goal was to make their research clear, engaging, and accessible for a general audience.
At the event, which followed Mila’s annual Partner Symposium, the six talented Mila researchers presented their research in three minutes to a jury and a full audience of peers and industry professionals.
They showcased a wide range of research areas at Mila, from medical imaging and AI-assisted urban design to robotics for people with disabilities and forest monitoring to tackle climate change.
The jury, hailing from BrainBoxAI, TELUS, Intel, Amgen and Mila, assessed the presentations based on public speaking skills, science popularization, structure of the presentation, and creativity.
This edition was made possible by the financial support of Mila partners BrainBoxAI, Telus, and Intel, who provided prizes and gifts for the winners and finalists.
Discover the presentations
1st place: Kusha Sareen, Master’s student at McGill University, for his presentation “Teaching AI to Check Its Own Math”
2nd place: Amar Kumar, PhD student at McGill University for his presentation “From Algorithm to Clinic: Explainable AI for Trustworthy Diagnosis”
3rd place: Rashid Mushkani, PhD Student at Université de Montréal, for his presentation “Locked Out by Design”
Audience Choice Award: Michael Przystupa, Research Intern at Université de Montréal, for his presentation “From Math to Mobility: Empowering People Through Smarter Control”
Carol Altimas, Master’s student at Université de Montréal: Becoming a Tree Detective
Benno Krojer, PhD student at McGil University : From a soup of raw pixels to abstract meaning
