Expandable microservices architecture for AI collaboration in real time.


Are you someone who is driven by the mission to create positive change for Albertans? Someone who values research excellence and the translation of AI-advancements to applications in financial services and society at large? We’re looking for a Managing Director to help guide our AI research teams and partnerships, and we’d love to hear from you.


In the spring we introduced you to the ATB AI Lab: ATB partnering with University of Alberta experts across the disciplines of science, mathematics, computer science, and economics pushing the boundaries of research and blurring the lines between theory and production-ready AI solutions.

ATB’s promise is to listen to Albertans and put them at the forefront of everything we do. We are striving to identify customer problems and to respond to opportunities the moment a customer need arises. To achieve this, one of the research streams at the ATB AI Lab is geared towards building a real-time Artificial Intelligence system. With vast amounts of data, we need to distinguish signals from an ocean of noise, so we are using big data techniques to process incoming data streams and we’re applying machine learning algorithms to extract informative behavioral signals. Make sure to subscribe to alphaBeta to get more details about  the complex problems that our research teams are tackling.

Through our partnership with UAlberta, we are able to invest resources into a flexible platform  instead of a single-purpose solution, that enables collaboration and will help us build for tomorrow. Decentralization and modularity are core to this model, so we’re creating an “extensible platform”: blended with architecture, language, and interfaces that will enable different components created by a team of researchers  to co-exist and interact.

We’re building an early prototype in a cloud-based infrastructure, focused on data storage, processing capability, machine learning algorithms, and APIs. Our microservices-oriented solution connects all the components of the system together, joining messaging buses, with mathematical models and databases to query data for real-time event detection. New components can be easily created and integrated into the overall system. The setup is like a ‘social-network’ for different components, allowing them to interact, and accommodate each other without significant dependence.

Why does industry-academia partnership benefit from a microservices solution?

Partnering with academic institutions provides a unique experience, in part because academic research tends to be horizontal in nature, with researchers working independently on different components of the system architecture. For the ATB AI Lab, building a microservices setup that is flexible, portable, and adaptable means that we can quickly change our tech stack when new challenges arise, without building an entirely new solution.

Rather than working separately and trying to integrate technical pieces at an mature project stage, we can add and remove components, include different databases, data transformation pipelines, and predictive modeling modules to react quickly to make sure we are at the forefront in solving problems for our customers.

Why start with a prototype of the microservices solution?

Our research partners are experts in their fields, but in some cases are not entirely familiar with banking. Launching a prototype phase gives teams a playground to get used to the types of data that we have and the types of problems we are solving in banking. It also provides us an opportunity to test the onboarding of new research teams and academic  partners—a sort of training phase before the wheels come off for both us and our research partners.

In addition to giving the researchers a sandbox, our in-house data science teams were able to see the immediate value of this prototype and are quickly deploying successful  solutions based on proofs-of-concept.

As we deepen our research partnership with UAlberta, we are excited to be building more than a single purpose technical solution - our aim is to have a system where components built by researchers, students, and ATB data scientists can interact seamlessly. With this system, we are translating the power of machine learning algorithms to applied banking solutions, to make banking better.