| Applicant | Co-Applicant(s) | Institution | Knowledge User Organization | Project Title |
| Athanasios Zovoilis | Sachin Katyal (KUC) | University of Manitoba | CancerCare Manitoba | AI4Omics-MB: A Manitoba Pilot of GenAI for Clinical Genomics |
| Abstract | ||||
| AI4Omics-MB is a Manitoba-based project that applies generative artificial intelligence (GenAI) to identify biomarkers using genomic data from the Manitoba Tumour Bank at CancerCare Manitoba (CCMB) and the associated Canada Precision Health Initiative (CPHI) PrairieGen Project. It adapts methods and principles aligned in AI4Omics, a Horizon Europe research initiative currently under review, in which the University of Manitoba and CCMB are active partners. Most AI tools used in biomedicine today are designed for text analysis and are poorly suited to genomic data. AI4Omics-MB addresses this gap by testing GenAI models that are trained directly on DNA and RNA sequences. These models can uncover patterns relevant to early cancer detection and treatment selection, while maintaining privacy and clinical relevance. The project will fund two full-time bioinformaticians and a project manager, with matched in-kind support from CCMB. The team will develop, deploy and validate AI models using local data, in collaboration with CCMB clinicians and academic researchers.This project aligns with Manitoba’s strategic priorities in emerging technology, interdisciplinary partnerships, and practical health innovation and will advance Manitoba’s capacity in AI-driven health research and support workforce development. | ||||
| Cyrus Shafai | Miyoung Suh (AC) Bruce Hardy (KUC) |
University of Manitoba | Ebb & Flow First Nation | Indigenous AI for Vision Resilience: Platform for Preventative Eye Health and Ethical Biotech |
| Abstract | ||||
| This project will establish Manitoba as a leader in Indigenous-governed artificial intelligence and bio-innovation by developing and validating an AI-enabled, bioactive hydrogel system to mitigate Computer Vision Syndrome (CVS) in Manitoba’s digital workforce. The initiative integrates predictive AI models, Indigenous-sourced carotenoids and wild-rice lipids, and ethical data stewardship aligned with OCAP and FAIR principles. Three primary objectives guide the work: (1) build supervised-learning models to identify occupational risk factors for screen-related strain; (2) design and validate hydrogel delivery systems optimized for sustained ocular protection, first-order release, and maintained antioxidant activity; and (3) embed Indigenous data sovereignty, digital governance, and benefit-sharing throughout research and commercialization. Experimental activities include AI-driven workforce exposure analytics; hydrogel synthesis and testing for release kinetics, stability, and antioxidant performance at the University of Manitoba; and Digital Twin pilots in St. François Xavier, Ebb & Flow First Nation, and Winnipeg workplaces. All work occurs within the Reimagining Food Systems Knowledge Nexus, ensuring Indigenous-led oversight, policy alignment, and community training. Outcomes include a hydrogel prototype, a Vision Risk Index powered by ethical AI, and a reproducible Indigenous governance framework. This initiative strengthens Manitoba’s AI, biomanufacturing, and Indigenous innovation capacity while supporting workplace health, economic reconciliation, and future commercialization pathways. | ||||
| Jay Toor | Aazad Abbas (AC) Edward Buchel (KUC) |
University of Manitoba | HSC Foundation | ORION – Operating Room Intelligence & Optimization Network |
| Abstract | ||||
| Manitoba faces a critical surgical backlog, with over 1,300 patients waiting and median wait times of 170 days—far beyond recommended benchmarks. Current operating room scheduling wastes resources: surgeons misjudge procedure times 75% of the time, and less than half of scheduled operating room time is spent performing surgery. This drives costly overtime, increases patient risk, and prevents backlogs from improving. This project will develop an intelligent scheduling system for Health Sciences Centre Winnipeg using machine learning to predict trauma surgery demand and procedure times, then create optimized schedules that maximize surgeries while minimizing overtime and improving outcomes. Unlike traditional approaches, our system learns to make predictions that produce the best schedules—not just accurate predictions. The system prioritizes healthcare equity by monitoring outcomes across demographic groups, including Indigenous and rural Manitobans. We expect to increase surgical capacity by 5-8%, reduce wait times by 10-15%, and save $2-4 million annually while building provincial expertise in healthcare innovation. | ||||
| Mina Nouredanesh | Ryan Zarychanski (AC) Nick Hajidiacos (KUC) |
University of Manitoba | University of Manitoba Departement of Internal Medicine (Max Rady College of Medicine) |
Transforming Manitoba’s Healthcare through AI-Powered Predictive Tools and Digital Twins Built on Three Decades of Medical Data |
| Abstract | ||||
| Acute and critical health conditions place a significant and growing burden on healthcare systems, with Intensive Care Units (ICUs) accounting for a disproportionate share of hospital costs and mortality. Yet, many patients at risk of deterioration are managed outside ICUs, highlighting the need for predictive and preventive tools that span the entire critical care continuum. This project aims to develop AI-powered predictive models and patient-specific digital twins using three decades of linked clinical data from Manitobans to enable precision care and optimize health system resource management. Leveraging foundation models pretrained on large-scale electronic health record datasets, these systems will be fine-tuned on Manitoba data to predict outcomes such as clinical deterioration, escalation of care, mortality, and ICU resource utilization. Predictive models will support early risk stratification in acute care, while digital twins, virtual, real-time simulations of patient physiology, will provide individualized decision support for critically ill patients in ICUs. Models will be validated for accuracy, interpretability, and fairness across diverse populations and integrated into clinical workflows through pilot testing. This initiative will enhance patient outcomes, improve efficiency in resource allocation, and strengthen data-driven data-driven precision care in Manitoba. | ||||
| Qian Liu | Enayon Taiwo (AC) Tobi Jolly (KUC) |
University of Winnipeg | Siloam Mission | From Absence to Action: A Multi-Agents AI Framework for Homelessness Service Disengagement |
| Abstract | ||||
| In Manitoba, Indigenous women, girls, and Two-Spirit people face much higher rates of violence and disappearances than other groups. Shelters and support agencies work very hard to keep people safe, but in large cities like Winnipeg it is difficult to notice when someone quietly stops using services. Staff cannot scan thousands of records every day, so people who are in danger may go unnoticed until it is too late. This project aims to help fill that gap. We use artificial intelligence (AI) “agents” to support frontline workers. Each AI agent is a tool that can watch data, predict risks, and send clear messages to staff. The system combines daily visit records, case notes, and official guidelines to spot when a super vulnerable person seems to have left services suddenly and may need quick follow-up. How it works? One agent predicts who might be at higher risk. Another agent checks whether a person’s absence is normal or concerning. A third agent turns this into easy-to-read alerts for staff. A final agent looks up policies and guidance to keep the advice accurate and up-todate. The goal is to help staff respond sooner and prevent people from slipping through the cracks. | ||||
| Rebecca Davis | Joshua Walsh (AC) Tomasz Kalkowski (KUC) |
University of Manitoba | Lithogen Inc. | AI-Driven Design of Lithium Sensors for Manitoba’s Critical Minerals Sector |
| Abstract | ||||
| Lithium is essential for batteries used in electric vehicles and renewable-energy storage. As Manitoba’s lithium industry grows, there is a need for fast, accurate, and affordable tools to measure lithium in complex samples found in the province. Traditional laboratory methods are accurate but slow, expensive, and not practical for use in remote field settings. This project will use artificial intelligence (AI) to accelerate the discovery of new organic molecules that change their optical properties when they bind lithium. An AI method called active learning will guide the search through thousands of possible molecular structures, continually improving its predictions as new computational and experimental data are generated. This AI-driven approach greatly reduces the time and resources needed to identify high-performing lithium-responsive molecules. The most promising candidates will be synthesized, tested, and incorporated into an early-stage optical sensor prototype. Lithogen, a Manitoba-based geophysical instrumentation company, will support this work by providing representative samples and ensuring the sensor design reflects field needs. By integrating AI-driven molecular discovery with sensor chemistry and industry guidance, this project supports Manitoba’s capacity in emerging technologies while advancing innovation in the province’s critical minerals sector. | ||||
| Tim Rogalsky | Gordon Goldsborough (KUC) | Canadian Mennonite University | Manitoba Historical Society | Ethical AI for Manitoba: The MHS Agentic Chatbot Project |
| Abstract | ||||
| This project will create a trustworthy, Manitoba-built conversational AI system that helps people explore the province’s history using reliable, clearly cited information. In partnership with the Manitoba Historical Society (MHS) and Tichon Technologies, Canadian Mennonite University (CMU) will develop an agentic retrieval-augmented generation (RAG) chatbot that answers questions about historic sites using verified data from the Historic Sites of Manitoba database. Unlike commercial AI tools, the system will be locally hosted, transparent about what it knows, and explicit when information is uncertain or unavailable. The research will combine technical development with user-focused evaluation. The chatbot will retrieve information from a secure CMU server, check its answers internally before responding, and provide citations for every fact it shares. New capabilities, such as showing connections among people, places, and events, or identifying nearby historic sites, will make Manitobas heritage easier to explore. The project will also strengthen provincial capacity by involving undergraduate students in applied AI research and ethical evaluation. Expected outcomes include a working public prototype, evidence on trustworthy AI design, open-source tools other Manitoba organizations can reuse, and a stronger foundation for ethical, locally governed AI across the province. |
||||
