Projets novateurs réalisés

Pragmatic Investigation of Artificial Intelligence for Symptomatic Venous Thromboembolism Prevention: PRISM-VTE Study

Venous thromboembolism (VTE), which includes deep vein thrombosis and pulmonary embolism, is a common and potentially life-threatening condition affecting 1 in 10 people during their lifetime. Many cases occur during or shortly after hospitalization. Although hospitals aim to prevent these events, accurately identifying VTE using routine hospital data remains challenging, as diagnostic codes alone are often inaccurate.

This project will develop and validate a more reliable way to identify VTE using a combination of diagnostic codes, medical imaging records, and treatment information from The Ottawa Hospital Data Warehouse. Suspected cases from 2023–2024 will be reviewed and compared with expert clinical adjudication to determine how accurately the algorithm detects true VTE events.

By improving VTE identification, this research will support better prevention strategies, enhance hospital quality monitoring, reduce unnecessary treatment, and strengthen patient safety. The project will also train undergraduate students in applied clinical research and health data analytics.

Advancing Reliability and Training Systems for On-Demand Transit – Argo and York University

Argo Transit Corporation provides municipalities with a turnkey, near door-to-door on-demand transit service that integrates software, onboard hardware, vehicles, and operational procedures. As Argo expands into more communities, the organization must ensure that service delivery remains reliable, consistent, and efficient across diverse operating environments. This project supports that goal by developing standardized processes, tools, and training systems that strengthen operational readiness and improve the quality of transit service delivery.

Interns will work across engineering, operations, and software development to create a repeatable vehicle readiness standard, including pre-delivery inspection procedures and installation instructions. They will also design operational playbooks, driver training materials, and internal tooling that helps Fleet Operators follow consistent workflows. Finally, the team will build software features—such as dashboard guardrails and a driver-app training mode—that reinforce Standard Operating Procedures and reduce operational errors.

By improving readiness, training, and operational tooling, this project helps municipalities deploy reliable shared transit more quickly and sustainably. These improvements support broader Canadian goals related to mobility, congestion reduction, and access to essential services, while providing interns with hands-on experience in engineering, operations, and real-world problem solving.

AI Enablement Summer Internship (4 months) at Good & Well – Sadia Tasnim

Good & Well is a boutique impact investment firm based in Toronto, Canada that invests in and supports inspiring entrepreneurs driving social change.

We are seeking the help of a student with relevant data science and/or computer science experience to support us in identifying areas of operations within our organization that could be streamlined, through automation and/or integration of AI tools.

This project would lead to positive economic and social benefits both for our organization, as well as for the 19 companies that are a part of our investment portfolio. Indeed, by streamlining and increasing efficiencies in our operations in an ethical and responsible manner, we notably aim to:
1) Increase our capacity to source companies who align with our investment approach and who could benefit from our financial and organizational support – which would help support Canada’s economy and strengthen the social entrepreneurship sector;
2) Strengthen our impact monitoring, measurement and reporting capacities, as well as those of our portfolio companies – which would support better social returns;
3) Elevate our capacity to identify potential partners and areas of collaboration that could lead to mutual benefit – supporting both economic and social growth.

Developing an AI-Enabled Smart Labour Market Decision Tool for Career Planning and Occupational Matching

This project aims to improve an online tool that helps people make career decisions. The tool will take a user’s resume and compare their skills, education, and experience to different jobs, showing how well they match and explaining the results in simple terms. It will also allow users to compare career options, see job market information like wages and demand, identify possible employers, and understand future risks such as automation. The goal is to have an easy-to-use app that turns complex labour market data into clear, personalized advice, helping job seekers, career counsellors, and other professionals make more informed choices.

Repenser la recommandation : un agent d’accompagnement pour la découverte de nouveaux contenus

À l’ère où les algorithmes de recommandation traditionnels agissent comme des « boîtes noires » prédatrices qui enferment les usagers dans des bulles de filtres, ce projet de recherche vise à redonner le plein pouvoir aux citoyens. En collaborant avec la plateforme de partage de connaissances « En commun », nous développerons un agent d’accompagnement à la découverte novateur, transparent et éthique. Cet outil permettra aux utilisateurs de formuler leurs véritables besoins et de choisir eux-mêmes leur parcours d’exploration afin de découvrir de nouvelles ressources et expertises. Pour la plateforme En commun et la société dans son ensemble, les retombées sont majeures : ce projet vise à concevoir une intelligence artificielle souveraine et respectueuse de la sobriété numérique. Il crée ainsi un espace numérique sain qui valorise le libre arbitre, protège la diversité des idées et renforce ainsi la résilience de nos écosystèmes communautaires et culturels.

Family Support Improvements

Since its inception in 2020, Makoon has assisted over 200 families and reunited 368 children with their families. As it continues to grow and support more families, it faces various challenges in its operation and delivery of family support services. Currently, some service delivery processes rely heavily on manual coordination and individualized staff approaches. While this flexibility benefits families, it creates inconsistencies in documentation practices, family progress tracking, and coordination between intake, case management, and family support roles. This project will help Makoon address challenges and issues that are not part of its day-to-day activities and will focus on improving various processes such as:
– family support documentation tools and
– communication processes between different roles and family support processes.

Investigating the impact of grazing on the soil microbiome and belowground carbon cycling

Changing farming subsidies provides an opportunity for habitat restoration in Cumbria. So that resources are allocated effectively within these landscapes, restorative actions must be assessed for nature recovery and ecosystem service delivery. Given the history of herbivore overgrazing in Cumbria, evaluation of the impact of grazing exclusion on belowground biodiversity and carbon cycling is needed. This project aims to understand how Cumbrian soil microbiome diversity and community composition and soil organic matter (SOM) composition stores changes with grazing exclusion. I also aim to link these belowground indicators to assess changes in soil functions across grazing intensities. To do this, I will use soil samples collected in Cumbrian upland grazing exclusion plots and use amplicon-based metabarcoding to identify bacterial, archaeal, arthropodal and vertebrate diversity and use high-resolution mass spectrometry to classify extractable SOM compounds. This project combines the research interests of the Forest Ecology and Conservation group at the University of Cambridge, the motivations of the Centre for Landscape Regeneration collaboration and the expertise, interests and methodologies of the Ecosystems and Global Change Group at Trent University.

Advancing AI Innovation Through Cross-Functional Collaboration (4)

AltaML builds artificial intelligence (AI)-enabled solutions to business problems. We work with organisations, bringing together their data and domain expertise with our AI expertise, to develop AI solutions that are deployed in their operations. We also commercialize AI-enabled products business via industry-specific ventures, yielding scalability from our investment in the first solution.
Wildfire workforce planning is under growing pressure due to increasingly severe and unpredictable fire seasons, rising burnout risk among frontline staff, and escalating overtime costs. In response, Marley will work alongside the Government of Alberta (GoA) to address these challenges through two core workstreams.
The first focuses on Duty Officer assignment optimization analyzing current scheduling practices, workload distribution, and assignment patterns to identify inefficiencies and develop smarter, more equitable approaches to rostering Duty Officers across fire seasons.
The second involves building predictive models and agentic AI tools that give leaders better foresight into workforce demand. This includes forecasting overtime requirements based on fire weather patterns and historical deployment data, flagging burnout and fatigue risk before it escalates, and surfacing actionable scheduling recommendations through intelligent decision-support agents that leaders can interact with directly.
Together, these workstreams aim to optimize pre-suppression overtime spending, support the health and sustainability of the wildfire workforce, and improve the quality of workforce planning decisions equipping GoA leaders with the insights and tools needed to plan proactively rather than react under pressure.

Advanced Optimizations for Quantum Compilation

Xanadu’s mission is to build quantum computers that are useful and available to people everywhere. Founded in 2009, with headquarters in Toronto, Canada, the company has grown to 250+ employees, many of them scientists, researchers and engineers, working across all disciplines (hardware, software, applications, algorithms) to achieve this goal. This project is within Xanadu’s software team, which is building the complete end-to-end software solution for quantum computing (QC).
A key component of the QC software stack is the quantum compiler, the software that transforms the highlevel user quantum program, to a low level machine code that can be run on the hardware – this is analogous to classical compilers targeting today’s traditional ICs (integrated circuits). The performance and size of the machine code is largely determined by the quality and sophistication of the compiler. In particular, optimizations implemented by the compiler are critical in ensuring that a high-level program is transformed to code that is as compact and efficient as possible. A program that is not compiled optimally, may completely squander the benefits of quantum computing, taking weeks instead of hours to run.
As Xanadu moves towards building a utility scale quantum computer, it is critical that the quantum programs themselves are built efficiently, with high optimization. This will ensure that real world problems can be tackled in reasonable time at reasonable cost. This project will study optimizations for quantum compilation aimed at improving efficiency. The work is expected to show a quantitative improvement in performance, through the use of advanced techniques within Xanadu’s compilation pipeline

Amélioration de modèle “Speech to text” pour les apprenants en langues secondes / Improvement of “Speech to text” model for second language learners

Nous avons un modèle entraîné sur des fichiers audios de personnes qui parlent le français en seconde langue, et nous souhaitons explorer et améliorer la reconnaissance de notre modèle “Speech to text” pour les apprenants débutants en français. Objectif: réduction du WER pour des apprenants débutants en français avec une IA conversationnelle.

Plus tard, nous voulons plusieurs l’adapter pour d’autres langues.

Le stagiaire analysera les performances du modèle STT existant, préparera et enrichira les données audio d’apprenants débutants, mènera des expérimentations pour améliorer la reconnaissance (fine-tuning, data augmentation, exploration de nouveaux modèles), évaluera systématiquement le WER et les erreurs types, puis documentera les méthodes, résultats et recommandations tout au long du projet.
Les livrables incluront un rapport d’analyse initiale du modèle, un corpus nettoyé et annoté, plusieurs versions de modèles améliorés accompagnés de benchmarks comparatifs, une analyse claire des erreurs récurrentes, ainsi qu’une documentation finale décrivant le pipeline, les méthodes testées et les meilleures pratiques pour l’amélioration du STT pour apprenants débutants.

Development of monolithic unitary operations for quantum photonics.

This project is based on a collaboration between the University of Ottawa and Wroclaw University of Science and Technology, aims to revolutionize how we control light for the next generation of quantum computers. Currently, managing light particles requires bulky and complex networks of mirrors and lenses that often leak information and are difficult to scale up. Our team is replacing these fragile setups with volume holograms, which are a compact, single-block glass elements that can redirect light with incredible precision and minimal loss. By using these holograms to perform complex calculations and even mimic long travel distances in a tiny physical space, we are creating a more efficient toolkit for high-speed quantum processing. This partnership benefits both institutions by merging international expertise to solve fundamental engineering hurdles, ultimately positioning both Canada and Poland as leaders in the global quantum race and directly contributing to the goals of Canada’s National Quantum Strategy.

L’expérience transformationnelle comme levier de fidélisation, d’implication et d’attractivité internationale pour MUTEK

Ce projet d’intervention en recherche appliquée s’articule autour de l’optimisation stratégique du modèle d’affaires de MUTEK par le prisme des expériences transformationnelles. L’étude vise à décrypter les mécanismes psychologiques et décisionnels des participants (tels que l’immersion sensorielle, l’innovation artistique et la qualité des interactions sociales) afin de comprendre comment ces leviers façonnent des attitudes positives fortes et durables. En mobilisant les concepts de marketing d’engagement et de participation active, la stagiaire développera une architecture de l’offre capable de convertir la résonance émotionnelle des expériences vécues en changements de comportements, notamment la promotion organique, le mécénat et la co-création de valeur au sein de l’écosystème de l’organisation.
Le projet de recherche a comme objectifs d’accroître la fréquentation touristique internationale; de pérenniser l’expérience au-delà du format événementiel éphémère; et de structurer un écosystème événementiel durable autour des activités de MUTEK.