Andrea Yzeiri serves on the board of WEtech Alliance and chairs the McGill MMA Advisory Board, drawing on her more than a decade of experience as a data specialist across sectors. (PHOTO COURTESY OF ANDREA YZIERI / University of Windsor)
By John-Paul Bonadonna
Andrea Yzeiri was supposed to become a lawyer.
As an undergraduate at the University of Windsor, she was pursuing two degrees with an eventual career in intellectual property or commercial law in mind. Then, somewhere along the way, she discovered statistics, optimization and research — and found herself increasingly drawn away from the career she had imagined.
“I kind of stumbled on it,” Yzeiri says of the path that eventually led her into artificial intelligence.
“I really fell in love with data.”
Today, that detour has taken her to the forefront of a rapidly changing field.
Yzeiri is chief data and analytics officer and lead AI engineer at Picsume, a Windsor-Essex SaaS platform for talent acquisition technology.
She also teaches graduate workshops in Python at McGill University, consults with organizations on business strategy and data, serves on boards, maintains ties to the University of Windsor and travels across Canada as a speaker on AI.
In September, her growing influence in the field was recognized nationally when she was named one of the 2026 recipients of the Women’s Executive Network’s Canada’s Most Powerful Women: Top 100, in the AI and Technology category — which recognizes women whose work moves technology from theory into real-world applications.
For Yzeiri, the recognition came earlier than she expected.
Finding a passion in research
The beginnings of Yzeiri’s career can be traced to her time at UWindsor, where she credits the combination of academic freedom, research opportunities and supportive professors with changing her plans.
She entered the Outstanding Scholars program and began working with Odette School of Business professor Dr. Fazle Baki, researching management science and operations management and applying analytical techniques to efficiency questions in the healthcare sector.
The experience quickly changed how she saw her own abilities and interests.
"I was very interested in management science, specifically the concept of decision modelling," she recalls.
"It was fun for me to envision, ‘OK, how do I build some kind of programming model that can solve this problem with these different variables?’"
Her undergraduate research eventually led to a thesis examining efficiency optimization in health care.
She also became the first woman in the Odette School of Business to complete an undergraduate thesis and went on to present her work at international conferences.
“I had so many opportunities at the University of Windsor that I'm not sure would have happened for me at another university,” she says.
“The exposure to research really opened my eyes.”
The experience also gave her a foundation she would draw on later at McGill, where she pursued graduate studies and encountered machine learning, modelling and algorithmic design.
It was there that her path toward AI began to take shape.
From models to responsibility
Yzeiri remembers being introduced to early versions of OpenAI's GPT models in her advanced machine learning class around 2021, when artificial intelligence was still far less accessible to the public than it is today. Her academic work exposed her to both building models and thinking about what they could do.
As demand for AI skills grew, so did opportunities to apply what she had learned.
Yzeiri says her interest in responsible AI emerged alongside her technical work rather than as an afterthought.
Her background in optimization had taught her to look for ways to make a model better. AI introduced a more complicated reality: there is no such thing as a perfect model.
“Humans are ultimately the ones creating these models,” she says.
“There's always issues that you're going to find with data as we’re inherently imperfect.”
Those issues can include different forms of bias introduced during the development and deployment of a system. Yzeiri is particularly interested in what she calls operational bias — considering not only how a system is designed, but also how people will actually use it.
“How are your end users going to be using your product?” she asks.
“Could they use it wrong? What does ‘wrong’ mean?”
That question has become central to her professional work. At Picsume, she leads data and analytics while contributing to the development of AI systems designed to support hiring and workforce decisions. The company describes her role as designing intelligent systems that turn complex hiring data into clearer, fairer, more actionable insights.
Her work is part of a broader career that has crossed technology, health care, academia, consulting and research. Yzeiri serves on the board of WEtech Alliance and chairs the McGill MMA Advisory Board, drawing on her more than a decade of experience as a data specialist across sectors.
A more measured view of the AI revolution
Despite working directly in AI, Yzeiri is notably cautious about some of the grander predictions surrounding it.
She believes AI is here to stay, but expects the current wave of enthusiasm to give way to a more pragmatic period in which organizations become more selective about where the technology actually makes sense.
“I think that we're going to go into another AI winter,” she says.
That does not mean she expects AI to disappear. Rather, she anticipates greater use of automation and traditional machine learning as organizations gain a clearer understanding of what newer AI systems can — and cannot — reliably do.
One reason is fundamental: neural networks, which are the base model for AI, are probabilistic rather than deterministic. While they can produce useful results, they are likely to produce errors, what we call hallucinations, and within AI systems these errors can compound.
For Yzeiri, that distinction matters enormously when organizations decide where AI belongs.
“We cannot use it for things where we need more deterministic outcomes,” she says.
That measured approach also informs the way she talks about AI itself. She is reluctant to describe a model as “smart” or imply that it possesses human-like intelligence.
“It’s math, it’s code,” she says.
Making room for the next generation
The recognition from the Women’s Executive Network has added another dimension to Yzeiri’s work: visibility.
She has spent much of her career in rooms where she was the only woman, particularly when conversations turned to highly technical subjects.
“I've seen and felt some sexism firsthand,” she says.
“Most of the rooms I'm in, believe it or not, it's usually all men.”
That can mean having to establish her expertise before being fully heard, an experience she wants to make less familiar to the women who follow her.
“I really want to be that visibility where people say, yeah, I can definitely do this,” she says.
“If a man can do it, I can do it too — whether that's changing a tire or writing code.”
Her own career, after all, is a case study in the value of not following a predetermined path.
She did not set out to become an AI leader. She followed an interest in research, discovered a passion for data, pursued new opportunities and allowed her direction to change.
That is also the advice she offers students considering careers in AI.
“Stay flexible,” she says.
“We've seen so many changes happening and sometimes you have to be willing to maybe try things that go against the grain.”
The technology will change. The jobs will change. The skills in demand will change.
Yzeiri’s advice is to change with them.
“Don't be too rigid,” she says.
“Be willing to go down different roads and make a path for yourself.”
For someone who entered university expecting to become a lawyer, it is advice that may be as much autobiography as career guidance.
This profile is part of a special series celebrating Alumni Week. Discover events and more at uwindsor.ca/alumni.
