Vector Faculty Member and Canada CIFAR AI Chair Graham Taylor is Next AI’s Academic Director
By Ian Gormely
October 19, 2020
Jennifer Arnold had enjoyed a successful two decade run in the financial and entertainment industries when she decided to set out on her own. In 2018 she co-founded MinervaAI, a regtech company that uses natural language processing and computer vision to automate data collection for risk management analysis. But after building a minimum viable product, she and her co-founders decided they needed help if they wanted to grow their company.
“When you go into a business, you’re a beneficiary of all this work that went before you. There’s already all of this infrastructure,” she says. “But as a startup, you are starting from the ground up.”
Running a successful startup can test even seasoned entrepreneurs and the best recognize when they need to look outside of their organization for help. This is why, despite their collective experience, the MinervaAI team turned to Next AI to help them navigate the unique startup business landscape they were entering.
Launched in 2017, the Next AI accelerator program was born out of entrepreneurship initiative NEXT Canada with a goal of growing and supporting the Toronto AI startup ecosystem. “It’s like a mini MBA getting rammed into your head over the course of four months,” she jokes.
After successfully navigating the application process (about 20 teams are chosen from more than 300 applicants) co-founders are split into business and technical tracks, depending on their strengths and weaknesses. Leaders on the technical side are expected to bring a certain level of technical prowess with them. “Building a successful AI startup requires taking AI research, making it work in practice, and scaling it up,” says Graham Taylor, Next AI’s Academic Director on the tech side and a Canada CIFAR AI Chair and Faculty Member at the Vector Institute. “So they have to have the technical competency to do that.”
Meanwhile, leaders following the business track, where the curriculum is organized by NEXT co-founder, professor at the University of Toronto’s Rotman School of Management, and Vector Faculty Affiliate Ajay Agrawal, are expected to have some notable past contributions to academics, industry, or unique life experiences.
The business courses look at everything from the overall landscape to trade regulations and case studies, says Patricia Thaine, co-founder of Private AI, an unstructured data de-identification company, and a Vector Postgraduate Affiliate completing a PhD on privacy-preserving natural language and speech processing at the University of Toronto. Thaine brought with her a rock solid technical background, but wanted a better grounding in business. “It gives you an awareness of what to look at and hones your intuition around how to make decisions.”
NEXT Canada was created with an eye to commercializing Canadian research. “There was a lot of investment going to the US,” recalls Taylor. “We certainly needed some acceleration happening in Canada.”
NEXT Canada helped launch a number of Canadian companies, including Vector Bronze sponsor, Dessa as well as Nymi, founded by the current head of Borealis AI Foteini Agrafioti. So expanding the principles of the original NEXT Canada program to AI startups – identify talented entrepreneurs and give them the support they need to build a business in Canada – was a natural extension. AI is a field where Canada has long held an advantage in the world. “In terms of the science and the know how and the talent, it’s something where Canada really punches above its weight,” says Taylor.
Next AI also takes the NEXT Canada program a step further. Where Next 36 looks at Canadian undergrads and recent graduates, Next AI accepts applications from people who already have extensive work experience. The program, which boasts the teams behind startups Senso, Babbly, Feroot among its alumni, is also open to international entrepreneurs, provided they build their businesses in Canada.
It’s funded by a generous group of sponsors, both companies and individuals keen to see Canada build on the existing pool of innovative talent. “These teams might be building something very useful, notes Taylor. “A lot of them end up having their first clients be the sponsor companies.”
As well as organizing the technical curriculum and tapping potential faculty (Vector Faculty Members and accomplished computer vision experts Sanja Fidler and Raquel Urtasun, have both taught Next AI courses), Taylor is also part of the committee that evaluates each year’s applications.
He was impressed by Private AI’s de-identification suite which integrates text and image de-identification software with just three lines of code, as well as their technical savvy. “I liked that they picked a very specific problem that seems like people could use,” he says. “Pieter Luitjens, their CTO, he’s got, embedded systems knowledge where he can actually build a very effective variant of the algorithm that will operate on your phone. That implementation-level problem kind of slows down other competitors.”
Meanwhile MinervaAI caught his interest with its founders’ unique mix of personalities and experience. “They seem to have this very interesting, but effective team dynamic,” he says. “Jen and her co-founder Victor Tay, they’ve had careers in banking and domain expertise. They’ve got great connections and so if they’re going to go back and sell the tech to the banks, that’s good to have that experience.” The third member of the team, Damian Tran meanwhile just finished his master’s and complements their business experience. “He’s really knowledgeable, he’s really enthusiastic, and he seems to have the ability to just really make things work on the practical side. So it’s a really unusual team dynamic.”
One of the biggest benefits of the program is the sense of camaraderie that develops between participants. “The network that you’re getting is other founders who are going through the same problems, having the same issues that you are,” says Thaine who’s made a number of friends through Next AI over the past four months. Arnold agrees. “It’s an entirely different ecosystem than our typical business world in AI.”
Since beginning the program in March, PrivateAI raised a pre-seed round of funding with an eye to raising a seed round sometime in the next year, while MinervaAI received a $100,000 pre-seed money from the accelerator Acceleprise. Both credit the program with giving them the contacts and know-how to reach these milestones. “The AI ecosystem is like a big secret club,” says Arnold, “and Next AI is a great entry point into that world.”
Applications for Next AI’s Class of 2021 open October 19, 2020.