Dr. John Quilty in the Department of Civil and Environmental Engineering leads a graduate student research team building artificial intelligence (AI) models that help communities better anticipate and respond to water-related risks.
The research is supported by the Sinnathamby Professorship in AI for Sustainable Solutions in the Faculty of Engineering, established by University of Waterloo alum Thiru Sinnathamby (BASc ’01, electrical engineering) with an investment of $500,000.
Sinnathamby, Vice-President of Engineering at NVIDIA, is passionate about helping to improve water quality, accessibility and safety — his family continues to work with various non-profit organizations including Charity: Water, to help fund wells across Africa and Asia. He also sees AI as a tool that can be used to push innovation further and do good.
“I am determined to demonstrate all the positive outcomes with AI,” Sinnathamby said. “Its ability to help us do better is real but how we take advantage of it is up to us. When the opportunity came to invest in research that uses AI to improve people’s lives, and with a focus on water, I jumped right in.”
Quilty’s research team is studying how water moves from source to tap, forecasting river flow, optimizing reservoir operations and predicting contamination in drinking water systems. A central part of the work is improving predictions for rare, high-stakes events like 50 to 100-year floods with results already feeding into a flood forecasting partnership with Natural Resources Canada. This could eventually extend to conservation authorities and municipalities without large datasets of their own that would benefit from shared, data-rich models.
“Predictive models often perform worst when the risk is highest,” said Quilty. “We’re training machine learning models to give decision-makers a trustworthy range of outcomes, not just a single forecast that may lack precision. This makes for better estimates of risk and improved disaster response planning.”
Thiru Sinnathamby and Dr. John Quilty.
Sinnathamby was drawn to Quilty’s research for good reason but was particularly taken with the calibre of the students who wanted to carry the work forward.
“Who you have on board is essential to your research,” Quilty said. “But to hire a research team requires resources. The professorship made it possible for me to recruit students with a passion for applying AI to water problems. The result is a strong team dedicated to making big impacts with AI in the water field.”
Soheil Khosravi, a PhD student in Quilty's lab, is reimagining how reservoirs are operated — the balancing act between flood control, drinking water supply, hydropower, irrigation and recreation that determines how much water gets released, and when. Many AI models in this field take a black box approach where data goes in and a number comes out, but the reasoning is opaque. Khosravi's approach is physically informed instead — it incorporates the reservoir’s physical dynamics and operating constraints directly into the AI framework. Release decisions are guided not only by patterns in historical data, but also by the physical behavior and operating limits of the reservoir.
That distinction is the practical payoff for the water managers who must act on the model's output. Someone weighing a release under flood risk needs more than an accurate number — they need to see why it's the right number. "What we're doing improves how these models perform because we're building physical principles into the AI,” Khosravi said. “New methods like this take time for practitioners to trust and it's our responsibility to earn that trust."
Sinnathamby sees this gift going beyond its research impact. As a student who benefitted hugely from his co-op experiences, he wants to pay it forward and encourage others through his own actions. He calls it the multiplier effect — if a student knows an alum invested in their work, and one day does the same for someone else, the return compounds through people, not patents. Quilty sees his own mentorship the same way, crediting an earlier supervisor for showing him what he now tries to give back to his own students. For Khosravi, the idea is simpler still. “Being kind,” he said, “is in our control.”
The research team: Deepak Meena, Sydney Buis, Issam Souassi, Dr. John Quilty & Soheil Khosravi.