Graduate researcher and CIBC’s Enterprise AI team advance quantum methods to predict rare events more effectively
Project at a Glance: When banks try to spot unusual transactions, the warning signs can be easy to miss, especially when suspicious activity is rare and constantly changing. During his internship with CIBC’s Enterprise AI team, MScAC graduate researcher Jay Patel took on this challenge by exploring how emerging quantum technologies could help. Guided by Erin Li, Director and Head of AI Research at CIBC, and academic supervisor Professor Xiaofei Shi, Patel developed new tools to more easily identify irregular transactions, helping banks respond more quickly and accurately.
When Quantum Meets Finance: The Predictive Challenge
Large financial institutions process millions of transactions daily, all of which require real-time monitoring for signs of unusual activities with high accuracy and low latency. Event data often exhibits severe class imbalance, with rare events accounting for less than 1 per cent of all transactions, which poses a challenge for machine learning models.
CIBC’s Enterprise AI team sought to determine whether quantum computing, with its ability to represent information in exponentially larger spaces, could provide richer representations of transaction data.
The goal was to evaluate whether quantum-enhanced methods could match or exceed classical methods while offering new insights into high-dimensional feature relationships.
Academic Expertise Meets Industry Innovation
MScAC graduate student Jay Patel (Computer Science concentration) brought a strong foundation in applied machine learning and computational physics to CIBC, shaped by his undergraduate studies at Pandit Deendayal Energy University and the MScAC program’s emphasis on bridging research and industry practice.
Patel joined CIBC for its leadership in adopting cutting-edge technologies and for the opportunity to apply quantum machine learning to real-world challenges. The MScAC program prepared him for this opportunity through coursework in quantum algorithms, machine learning, and deep learning, as well as applied research training and mentorship.
“Turning ideas into practical finance solutions at CIBC showed me what it truly takes to build at scale and will help shape how I approach my future research in quantum computing for real-world problems,” said Patel.
His focus on quantum algorithms and optimization grew over four years of independent study and applied work, alongside achievements such as global quantum hackathon prizes, a CERN internship in QML for high-energy physics, and co-founding Bloq Quantum to support enterprise quantum software development.
Transforming Research into Impact
The research identified trade-offs between model performance, scalability and hardware constraints, providing actionable insights for CIBC’s AI research roadmap.
Patel’s comparative analysis — evaluating on-premises quantum simulation, GPU acceleration and ensemble quantum-classical strategies — helped shape ongoing exploration into hybrid AI models for various predictive tasks.
“We value the academic partnerships with MScAC for the fresh perspectives and unique expertise students like Jay bring to our ongoing research initiatives,” said Li.
A Culture of Innovation and Mentorship
CIBC’s Enterprise AI team fosters a culture of exploration, where applied research drives real business outcomes.
“Continuing to evolve our AI capabilities relies on top-tier talent who think outside the box,” said Ozge Yeloglu, VP of enterprise advanced analytics and AI. “Collaborating with academic institutions like the University of Toronto enables CIBC to bridge the gap between theory and practice, strengthening its leadership in applying AI to the financial sector.”
This mentorship model reflects the MScAC program’s mission: cultivating technically advanced professionals capable of bridging research and enterprise.
What This Means for Quantum in Banking
This project illustrates the emerging potential of quantum machine learning in financial applications.
By designing learnable, optimized quantum feature maps and benchmarking them against established machine learning systems, Patel contributed to both CIBC’s innovation pipeline and to the broader field of quantum algorithm design for real-world data.
Future opportunities include exploring ensemble models that combine classical and quantum learning, and scaling quantum kernel computation for larger datasets.
CIBC is committed to deepening its collaboration with academic partners to further advance applied research. Additionally, CIBC plans to scale AI and quantum capabilities across the enterprise, integrating research outcomes into long-term strategic initiatives.
By the Numbers
- Partner: Canadian Imperial Bank of Commerce (CIBC), Toronto
- Data scale: Quantum model trained on data scale not observed in previous studies
- Techniques: Quantum feature map, quantum machine learning, GPU-accelerated simulation
- Supervisors: Erin Li (CIBC), Professor Xiaofei Shi (University of Toronto)
Contact: For media enquiries, please contact MScAC Partnerships at partners@mscac.utoronto.ca. For more information about artificial intelligence at CIBC, visit www.cibc.com.
