When war disrupted her studies in Ukraine, Kseniia Buzko faced an uncertain future.
Determined not to let her education be derailed, she built a new future in Canada - earning top grades, forging international research collaborations and producing award-winning work in sports analytics at the University of Waterloo.
A new beginning in Canada
Buzko's journey to Canada began in 2022, when studying at her home institution in Ukraine became untenable. She applied and was accepted into the first cohort of the summer program for students from Ukraine whose studies had been disrupted by the war, a joint initiative between the Waterloo and the University of Toronto. Buzko spent her first summer at the University of Toronto, and through the program, began to build early connections with the Waterloo community - an experience she describes as life-changing.
"The summer program gave me the chance to resume my education after the war disrupted my life," says Buzko. "It offered academic stability, helped me refocus on my future and introduced what could be possible for me in the Canadian academic system."
She later transferred to the Waterloo to complete her final undergraduate year in computer science, graduating with top marks. An internship at Waterloo.AI followed, and under the guidance of Dr. Paul Fieguth, a systems design engineering professor and Waterloo's associate vice president - academic operations, she led a joint research project bridging her home university and Waterloo. That work became her undergraduate thesis and set the stage for her graduate research.
Inspired by that early momentum and a strong sense of belonging, Buzko knew Waterloo was the right place for her to continue her studies.
Breakthrough research in sports analytics
Now a master's student in the Department of Systems Design Engineering's Vision and Image Processing (VIP) Research Group - which is focused on visual processing, perception and responsible AI for real-world applications - Buzko is conducting research under the supervision of Professors David Clausi and Yuhao Chen focused on action recognition in ice hockey broadcasts.
Generic AI models often struggle with hockey, where players are smaller on screen, wear bulky gear and use sticks that standard systems don't account for.
To improve accuracy, Buzko developed a custom pipeline called CP-Hockey. It helps the AI system better interpret hockey footage by adjusting for differences in player size on screen and by adding important visual details - like head position and stick movement - that most models ignore. These improvements make it possible to tell apart similar actions, such as different skating directions, quick stops, stickwork and puck strikes.
"These contextual priors significantly improve recognition accuracy and enable more reliable analysis of game strategies and player performance," Buzko explains. "While developed for hockey, the methods show promise for other equipment-heavy sports as well."
In June, her work earned the Best Paper Award at the Linköping Hockey Analytics Conference for her publication Ice Hockey Action Recognition via Contextual Priors.
Kseniia Buzko with fellow members of the Vision and Image Processing (VIP) Research Group, where she develops AI models to improve action recognition in ice hockey.
Building a community
For Buzko, Waterloo has become much more than an academic home. It's also a place where she's found belonging and support.
"I love being part of the VIP Lab where I can engage in meaningful research and collaborate with brilliant peers and mentors," she says. "Beyond academics, I'm grateful to continue here my lifelong passion for playing badminton. It's helped me build close friendships that make me feel truly at home here."
She is equally grateful for the space that Waterloo has offered for her to heal and thrive.
"The environment at Waterloo supports not only my academic development but also my personal well-being," Buzko says. "I feel incredibly motivated here, and I look forward to continuing my journey at Waterloo as a PhD student."
