Even as a kid, Maksims Volkovs wasn't afraid of a challenge.
Growing up in Riga, Latvia as the son of a mathematics PhD and computer science professor, it's no surprise that Volkovs' interests gravitated to computer coding, solving advanced math problems and programming video games.
While he enjoyed sports like basketball and other hobbies, something about the prospect of overcoming a challenge, of using math to connect the dots that others couldn't see to help find the simplicity in something complex, kept pulling him back.
After his family immigrated to Canada in 2000, that same determination and passion for problem solving helped him learn English once the family settled in Toronto, enabling him to graduate from high school and earn acceptance to the University of Toronto where he enrolled in computer science.
During the time Volkovs was an undergrad, the technology industry looked far different than it does today. Smartphones were only a novelty, and the social media age was in its infancy. Artificial intelligence (AI) was regarded by many as closer to science fiction than reality.
But Volkovs wasn't interested in apps and phones. Always looking for a challenge, it wasn't long before he found his calling in a class entitled, "Introduction to Neural Networks and Machine Learning," in the winter of 2006.
In other words, AI.
"Back then, the field was completely dead," Volkovs said in an interview from his office at Layer 6, the Bank's AI division, in Toronto.
"But I could see the potential."
Today, Volkovs is the Chief AI Scientist at TD, where he oversees research and development at Layer 6, using AI technologies to design products and solutions to help usher in a new era of banking for TD colleagues and customers.
The challenge Volkovs and his team now face is building a world class AI lab that is focused on helping TD develop better ways to reduce complexity within the organization. At the same time, Volkovs believes there is an opportunity to make AI more understandable, approachable, and trustworthy for TD colleagues and customers, while providing personalized solutions that can help improve the banking experience at TD.
"AI is important to the TD strategy to help customers gain better insights into their finances, so that they can make the right decisions for them and their families," Volkovs said.
Of course, that's easier said than done.
The "winter" of AI's discontent
When he walked into that U of T classroom on that fateful day in January 2006 for his first neural networks course, Volkovs didn't realize the professor was going to change his life. Or, for that matter, the very nature of AI.
He saw hope in what he called the field's "winter." Skeptics saw a dead-end, arguing that there wouldn’t be enough computer power, data or efficiency to make it work. He saw the potential to build something from the ground up.
The professor on that day was none other than Geoffrey Hinton, the British-Canadian academic and future Nobel Laureate, who is now often referred to as "The Godfather of AI." Volkovs remembers watching Hinton teach and was struck by his infectious passion.
"For all the focus on the course material, what really stood out was his unwavering dedication to AI and enthusiasm for the possibilities," said Volkovs.
"I wanted to be part of that, and to make that potential real."
Despite the general darkness during the so-called winter for the field while the skeptical view was common, Volkovs said Toronto was a bright spot. Anchored by Hinton and eager students at U of T, researchers quickly established the city as a global hub for AI research.
Hinton and other researchers at U of T are credited with pioneering numerous AI technologies and breakthroughs in this time. There were fundamental advancements in deep learning, where computers analyzed complex patterns in a similar manner to the human brain, enabling machines to process far more data with greater accuracy.
Researchers at U of T were also responsible for a major breakthrough in Acoustic AI, which helped lead to the development of smart home assistants. And there was a giant leap forward in computer vision, which led to the creation of AlexNet, a neural network that recognizes images.
For the researchers and students who studied at U of T at the time, Volkovs said it was impossible not to be inspired by the leaps and bounds by which AI technology was progressing.
That inspiration led Volkovs, along with Jordan Jacobs and Tomi Poutanen, to co-found the AI company Layer 6 in 2011, which would eventually be acquired by TD in 2018. Layer 6 would become the Bank’s AI centre of excellence, responsible for developing and implementing AI solutions across TD.
Layer 6 and TD
Volkovs and his co-founders created Layer 6 to use AI to analyze large amounts of data towards being able to anticipate customer needs.
An early sign of the company's potential came through, of all things, a contest. Volkovs and two other Layer 6 colleagues won the prominent ACM RecSys Challenge in 2017, the leading international competition on recommender systems. They put forward a new approach to improve AI system recommendations from a "cold start," where not much information is known about a customer.
The next year, Volkovs and a team from Layer 6 won RecSys again, blowing away the competition by providing the best music recommendation algorithm, predicting what listeners would like to hear next.
TD was enthusiastic about Layer 6 and its focus on providing customer insights which TD believed could help it achieve a possible AI leadership position.
"Layer 6 adds new capabilities to the Bank’s growing base of innovation talent and know-how," said a TD executive at the time of the acquisition.
"Artificial Intelligence has the potential to power a new generation of data-driven applications from personalized and real-time advice to predictive analytics that will shape the future of banking for millions of individuals."
Today, the Layer 6 office is located just a few blocks from where Volkovs first attended class with Hinton, and the team now numbers more than 200 colleagues, including engineers, data scientists, product developers and more.
As organizations are racing to create new AI products and services, TD is investing in this future. Indeed, the Bank announced in April that Layer 6 would open an office in New York City later this year, aiming to initially hire 20 data scientists, applied machine learning scientists, and Generative AI implementation specialists.
"When we were starting out, it was a small but very ambitious team," said Volkovs, referring to the approximately 20 employees at Layer 6 when it was acquired.
"The goal was always to build a global centre of excellence where we can attract global talent and deploy AI on a global scale. The field is changing very rapidly, and we aim to lead that change."
To do so, Layer 6 has teams of scientists who specialize in different areas of AI. They experiment with the latest technology and advance it through their research.
"Based on that, we help formulate the view of where the Bank should invest at an enterprise level in AI and where we can produce the most value," he said, saying that part of the challenge is to identify where the Bank might be able to make its next leap forward for customers and colleagues.
Earlier this year the Bank had that kind of moment when it launched TD AI Prism – a cutting-edge foundation model which can more accurately predict customer needs towards better personalized banking solutions.
Volkovs gets animated when he discusses the possibilities of TD AI Prism and what it means for the future, as the technology can scale much more efficiently to help provide more solutions with the same amount of resources.
In that way, it checks the boxes of the mandate to provide faster and more accessible solutions for customers and colleagues, building on the more than 60 AI solutions that the Bank has already implemented.
"I think TD AI Prism will drive the future of how our businesses operate by leveraging predictive AI to drive proactive decisioning at scale that was not possible before.”
Trustworthy AI
It's easy to think of AI as an abstract algorithm developed in a black box. But it comes from somewhere. It's the product of people, and people are imperfect, Volkovs said.
That's why, underpinning research Layer 6 is involved in is what's known as Trustworthy AI – the Bank's term for its award-winning team and approach to a responsible AI governance framework – in addition to meeting its legal requirements.
While the nuances of deep learning that forms the foundation of AI can be complicated for the average person, the mission that Volkovs and his team are focused on at the Bank can be summed up in a simple mantra: How can AI at the Bank make services and products simpler, more accessible, and faster for customers and colleagues?
For Volkovs, it starts with building solutions on a foundation of Trustworthy AI.
"People need to be able to trust their bank," he said.
That trust comes from transparency, Volkovs said. It's part of the reason why Volkovs promotes a culture that encourages colleagues to show and share their work. Layer 6 researchers are encouraged to submit their latest research papers to leading AI conferences, where they are peer reviewed by a community of experts. These papers are also publicly posted on the Layer 6 website, building transparency and trust.
Volkovs believes that building the right kind of AI requires more than considering what's technically possible. It also involves facing the challenge of building AI that people can trust.
That requires the right kind of team, that comes from a nurturing, and collaborative culture. That culture needs a mission too, and at Layer 6 it is a relentless pursuit of scientific innovation to uplift customers and colleagues through better experiences.
Connecting the dots
Building relevant AI solutions that make for better customer and colleague experiences isn't easy.
But since he was a child programming his own video games, Volkovs hasn't been intimidated by the complicated math and technical obstacles. In fact, it's where he's comfortable, making complex challenges simpler, connecting the dots, and providing an elegant solution by working with the team.
His next challenge is bigger than math or a video game – it's about how to bring bank services at TD into a new era.
Doing so by making AI banking solutions that provide simple, more accessible and faster experiences for customers and colleagues is a journey that Volkovs looks forward to.
It may be a challenge, but he's always embraced those, realizing potential he can see by building one solution after another, and taking that passion he felt in the classroom to the real world.