Projets novateurs réalisés

Generative AI as a service for Small Business

Small and medium-sized businesses (SMBs) face increasing pressure to remain competitive in a rapidly evolving digital landscape. Many SMBs face obstacles in adopting AI solutions due to limited resources and the complexity of available products, which are often too generalized to meet the specific needs or scale of smaller businesses [1]. Existing AI solutions are typically designed for larger enterprises and lack the customization required for the unique operational challenges of smaller businesses. As a result, these companies are unable to efficiently utilize their data—such as customer service logs, website analytics, and social media feedback—to drive informed decision-making and improve customer engagement. To address this problem, there is a need for a solution that enables SMBs to transform their unstructured data into a usable, interactive knowledge base, empowering them to make data-informed decisions and strengthen client relationships. This requires an approach that is both accessible for non-technical users and capable of processing complex data in a way that delivers relevant and meaningful insights.
WizeWerks, a consultancy specializing in AI solutions, is focused on empowering SMBs to access the benefits of generative AI without the need for extensive in-house expertise. As a solution to this this, WizeWerks wants to develop a white-label platform that SMBs can brand as their own which will allow them to offer an AI-powered support agent that interacts with their unstructured data and transform such data into an interactive knowledge base. The system will generate meaningful insights to help SMBs respond to customer inquiries, predict emerging trends, and make data-informed business decisions.

Kanin Energy-Heat Pump Feasibility Study

The project will consist of conducting a feasibility study and a preliminary design on the application of heat pumps for heavy industry to achieve decarbonization and economic objectives. Heat pumps can play a significant role in helping heavy industries achieve decarbonization by efficiently capturing and utilizing low-temperature waste heat, upgrading it to higher temperatures, and providing thermal energy for various industrial processes.

Scalable AI-as-a-Service (AIaas) Platform Development

This project focuses on building the core infrastructure for Farpoint’s new AI-as-a-Service (AlaaS) platform. Two interns will work with Farpoint engineers to create a robust and scalable system using Kubernetes for managing resources and efficiently distributing workload. The interns will also implement load balancing and network routing to ensure smooth handling of high user traffic and AI model requests. This new platform will allow Farpoint to reduce reliance on costly third-party providers, give them greater control and flexibility in deploying and managing AI models, and ultimately, strengthen their position as a cutting-edge provider of innovative AI solutions.

Data Collection and Model Development for Prosthetic Fitting using a Robotic Gait Simulator

ProsFit is a medical device manufacturing company that supports prosthetic clinic owners to radically improve productivity and user outcomes. Through innovation, ProsFit strives to provide limb wearers a choice of affordable, reliable and desirable prosthetic products and services. ProsFit has developed and is commercializing the solutions to provide improved mobility and quality of life to millions of amputees, at scale, and has attained significant professional usage (>700 amputees fitted worldwide). The proposed project will conduct collaborative research with the Centre for Intelligent Manufacturing (CIM) at Sheridan College with the aim to gather data for enhancing the comfort of prosthetic users with lower-limb amputations. The challenge is that a large dataset is required to improve provision services and outcomes. Through CIM, the project utilizes a 6-axis robot set-up developed at CIM that can mimic human movement patterns, allowing for thorough testing of how changes in variables affect force distribution and other data through our Robotic Gait Simulator (RGS). Long-term engagement between ProsFit and CIM is planned to develop this technology, and the proposed project will establish data collection capabilities on the robotic gait simulator, engage in data collection, and begin to build the model for prosthetic fitting with best outcomes.

Community Prosperity Hub & Non-Profit Data Collaboration

The Greater Fredericton Social Innovation (GFSI) team and its partners have developed an open-access data repository called the Community Prosperity Hub, launched in 2023. This Hub is still in development but to-date it serves as a powerful tool for nonprofits in the Greater Fredericton area by making socioeconomic data accessible through user-friendly visualizations, summaries, and resources. These tools help organizations understand their impact, improve service delivery, and secure sustainable funding. Aligned with the 17 United Nations Sustainable Development Goals (SDGs), the Hub positions GFSI as a leader in community-level data initiatives.

The Hub has already reached over 300 stakeholders, including United Nations representatives, demonstrating its potential for broader use and adaptability. To further develop this platform, GFSI seeks to engage a computer science graduate to incorporate additional social determinants of health and link each Hub with Canadian funding sources.

GFSI is also advancing social inclusion initiatives by supporting the City of Fredericton’s Anti-racism Task Force recommendations. This includes developing the Canadian Atlas of Social Inclusion—an online tool for reporting incidents of racism—as well as the Fredericton Venues application, which enhances access to community spaces. These efforts showcase GFSI’s commitment to fostering an inclusive and data-driven community.

Conception et développement d’outils automatisées pourcaractériser, modéliser et quantifier la morphologie dessystèmes ostéo-articulaires en trois dimensions à l’aide deplusieurs vues radiographiques

Les problèmes de douleur chronique reliés à des troubles musculosquelettiques affectent une grande partie de la population et les coûts directs (visite médicale, traitements, médicaments) et indirects (absentéisme au travail, productivité réduite) sur le système de santé sont extrêmement élevés. Pour améliorer le diagnostic, la planification chirurgicale et le suivi des pathologies de la colonne vertébrale, de la hanche et du genou, il est essentiel de considérer et de visualiser les articulations et les paramètres cliniques qui leurs sont associés en trois dimensions. Ce projet vise à concevoir, développer et transférer des outils informatisés innovateurs afin d’améliorer et d’enrichir les plateformes d’imagerie d’EOS Imaging et de Global Imaging On Line, notamment dans le but d’optimiser la reconstruction 3D, l’analyse morphologique 3D des os et assurer une prise en charge personnalisée des patients

Integration of Inflation Signal in Picton Mahoney’s Asset Allocation Models

Picton Mahoney Asset Management (“PMAM”) was founded in 2004 to provide unique investment solutions to institutional, retail and high net worth investors in Canada and around the world.

The Multi-Strategy Team at PMAM uses economic cycle models to determine asset allocation preferences across time during different macroeconomic regimes. We have recently incorporated an additional inflation signal to our cycle models which requires more work to integrate into the asset allocation process. We have identified other ends uses, including our 40/30/30 marketing campaign and our new Wealth Management unit. We also require automation of reports that will alert the multi-strategy fund managers of any new signals that would necessitate a change in asset allocation.

University of Toronto’s Master of Mathematical Finance (MMF) students bring highly marketable and specific skill sets that can support PMAM’s analysts and portfolio managers within the Multi-Strategy and Macroeconomic Research Teams. Their Matlab, Python and SQL programing skills will move our projects forward by allowing us to update and automate our existing cycle models that make asset allocation decisions.
These skills will also allow us to expand the models to new use cases within the firm, including the new Wealth Management unit. MMF students have basic knowledge of key concepts such as Portfolio Construction, Asset Allocation, Factor Investing and the Global Macro investing environment. This allows MMF grads to be quickly integrated, making an immediate impact. They can also design, generate and automate reports related to the operation and monitoring of Multi-Asset and Multi- Strategy portfolios.
Providing MMF students with the opportunity to integrate their academic studies with hands-on experience at PMAM allows us to support the intern’s journey from academia to industry, while benefitting from their intellectually curious mindset, their ability to think differently, and their aspirations provide unique investment solutions to Canadian and global investors.

AI-Powered Course and Lesson Generation Platform

Our project aims to transform real estate training through an innovative AI-powered platform that generates customized course outlines and lessons. Designed for real estate teams and brokerages, this platform enables team leads to prompt AI for training material tailored to specific skills and market demands. By incorporating BASL’s data on team insights and real-world examples, our solution ensures agents receive relevant, on-demand training. The platform will also feature a conversational AI module, simulating client interactions for practical, scenario-based learning, enhancing agents’ readiness for real-world challenges.

Professional Networking and and Intelligent Referral System for the Real Estate Industry

Our project aims to revolutionize real estate networking by creating a professional platform exclusively for realtors across Canada and the United States. This dedicated space allows realtors to build their brand, showcase testimonials, and facilitate trusted referrals with fellow agents. Functioning similarly to LinkedIn but specifically designed for real estate, this platform empowers agents to establish a network of trusted connections for referrals across cities, strengthening collaboration and enhancing client satisfaction.

Developing diffusion models for wind downscaling

This research project aims to develop a ready-to-use model for predicting detailed wind patterns at wind farms to help the industry better plan turbine placements and forecast energy output. Wind speed data is usually available at a large, low-detail scale, which makes it hard to predict specific site conditions accurately. Generating high-resolution wind data using physical models is accurate but is not always practical due to high financial cost, while simpler approaches can lack accuracy. Our project will develop a model by using advanced machine learning models, particularly diffusion models, trained on high-quality simulated data. This approach will make it easier and cheaper to produce detailed wind data at a finer scale. The intern will work on developing these models by including relevant weather variables, using advanced metrics to ensure accuracy, and testing the models’ ability to generate realistic wind patterns.

Controlled Self-Injection Techniques for Enhanced Laser Dynamics in Silicon Photonic Systems

The growing demand for high-performance computing (HPC), artificial intelligence (AI), and cloud applications is pushing traditional semiconductor technologies to their limits. This project advances silicon photonics, which uses light for faster, more efficient, and scalable data transmission, making it critical for next-generation computing and communication systems.
Focusing on hybrid quantum light sources and nonlinear optical materials, the research aims to stabilize laser systems against destabilizing optical feedback. Specialized electro-optical circuits ensure reliable single-wavelength laser operation, vital for optical communication and quantum technologies.
Photonic Wire Bonding (PWB) is a key innovation, a precise, scalable method connecting optical components with polymer waveguides. This approach supports customized laser sources for data centers and quantum computing applications.
Conducted at the University of British Columbia, the project drives advancements in energy-efficient data transmission, reduced thermal loads, and reliable system performance. It bridges theoretical and practical breakthroughs while fostering collaboration between Taiwan’s semiconductor expertise and Canada’s cutting-edge research facilities. This work marks a step toward the widespread adoption of silicon photonics in HPC, AI, and quantum systems, enhancing innovation and international cooperation.

Study of the Large Passive Deformation of a Slender Flexible Wing Using Continuation Techniques

The performance and stability of flexible mechanical structures in fluid flow can be improved by imposing bending-torsion coupling. In these structures, a bending deformation induces a torsional deformation which can be beneficial for the performance and stability. For instance, flexible wings can be tailored in such a way to have large deformations and twist passively to decrease the angle of attack and the structural loading. To evaluate the performance of a flexible wing with bending-torsion coupling, the wing is modeled as a slender rod with wing cross section using Kirchhoff’s rod theory. A semi-empirical drag formulation is also used to evaluate the fluid loading. The governing equations are solved with continuation methods using existing software packages such as MANLAB or AUTO to investigate on the existence of the bifurcation and structural divergence. In addition, the aerodynamic performance of a flexible wing with large deformations is studied. The results of this project are presented as a journal paper and a seminar.