This is part 10 of The Logic’s in-depth series exploring how Canada is faring in the global competition for tech talent, as economies reopen and companies and governments jockey for advantage in a remote-work world. Read the rest of the series here.

OTTAWA — Sam Pasupalak describes it as “mission impossible.” In the early 2010s, he and his co-founders were trying to recruit machine-learning specialists for their AI startup, Maluuba, but “Google and Microsoft and Facebook could pay at least five times what I [was] able to pay,” he recalls. 

A decade on, Canada has many more such workers—but it also has even more startups, scale-ups, large corporations and multinationals seeking to employ them.

Canadian AI is booming, and pandemic-inspired automation is likely to ensure that boom continues. Companies raised US$2 billion in venture capital across 147 deals in the year through mid-November, double the previous full-year record haul of US$1 billion in 2019, according to data from Pitchbook. They’ll likely spend much of that on new hires.

Meanwhile, foreign tech giants continue to set up AI shops in major hubs across the country. Indian IT consulting firms HCl Technologies and Mphasis announced new facilities this year in Mississauga, Ont., and Calgary, respectively. Others have been here for a while, including Facebook and Unity Technologies in Montreal, Fujitsu in Vancouver, as well as Alphabet companies Google Brain in Toronto and Montreal and DeepMind in Edmonton.

Those firms are competing amongst themselves—and with companies in other countries—for highly technical and skilled workers with machine learning (ML) and other AI expertise. “There’s only a handful of those people in the world,” said Arif Virani, chief operating officer of Waterloo, Ont.-based DarwinAI.

RBC’s Borealis AI institute, launched in September 2016, now employs 120 people, the “vast majority” of whom are AI PhDs, said head Foteini Agrafioti. Borealis does fundamental research, but also applies it to the financial-services firm’s products, like its money-management app.

“We don’t need to position or try to sell Canada to international talent anymore,” said Agrafioti, also RBC’s chief science officer. “People are very aware that [this is] where you go to do top machine learning.” She attributed some of the shift to the $125-million federal Pan-Canadian AI Strategy, announced in March 2017, which has paid for more than 100 research chairs.

Domestic firms must still compete for talent with multinational tech firms’ AI shops. In a moment local founders frequently cite, then-Alphabet executive chair Eric Schmidt showered praise on the country’s AI prowess during a November 2017 conversation with Prime Minister Justin Trudeau at a Toronto event hosted by the search giant. “We are enormously thankful to Canadians for inventing all this stuff, because we now use it throughout our entire business, and it’s a major driver of our corporate success,” he said at the time. “So we owe you.”

“I can think of a couple that we’ve lost to some of those larger labs,” said DarwinAI’s Virani, whose technology helps clients build better models. Startups also face a size-specific disadvantage when recruiting specialist AI talent. “Pure ML people really want to work with big data,” said Mike Murchison, CEO of Toronto-based Ada. Startups typically haven’t accumulated information stores as large as those available to giant tech platforms.

Young Canadian firms that do manage to corral lots of AI PhDs can become takeover targets for multinationals. Pasupalak grew Maluuba in Waterloo and then Montreal to about 50 people, including 25 “doing some sort of machine learning.” The company was signing deals with top auto manufacturers for AI-powered hands-free systems when one of Seattle’s twin tech giants announced in January 2017 that it would acquire the startup. “Microsoft bought us for our technical team and talent,” said Pasupalak. Maluuba’s founders in turn saw an opportunity to embed their technology in many more products.

But executives at some of Canada’s most prominent AI scale-ups say you can’t build a successful company with just ML PhDs. They cite Montreal-based Element AI, which claimed to have over 100 such experts on staff and raised over US$250 million, including from the Quebec government. That hiring spree “created massive inflation” for such talent, according to Louis Têtu, CEO of Quebec City-based Coveo. “The only purpose [to] bring together so many data scientists is to sell the workforce.”

In November 2020, Santa Clara, Calif.-headquartered ServiceNow announced it was buying Element AI, in a deal that Quebec Economy Minister Pierre Fitzgibbon described as a wash for investors. “Probably a hundred companies could have taken advantage of [that] AI talent to improve their digital offering,” said Têtu. “It exported the value of all that great education.”

Coveo itself makes AI search and customer-service tools for businesses, and last month raised $215 million in an IPO on the Toronto Stock Exchange; the company went public in part so it could grant restricted stock units to employees, a key hiring incentive. It has more than 625 employees, about 260 of whom work in R&D. AI is “really the evolution of software—[it’s] a theme, not an industry,” said Têtu.

Building a company around it requires “a combination of talent across the digital spectrum”—software architects, cloud-ops developers, UX designers, product managers and so on. That broadens Coveo’s competition to anyone hiring tech talent in Canada.

Ada now has a workforce of nearly 500. About 15 are highly specialized ML experts, and Murchison plans to double that number by the end of next year. Ada, whose platform firms like Air Asia and Verizon use to run customer-service chatbots, isn’t trying to aggregate as many PhDs as possible. “We’re focused on making this easy,” he said. “There’s a lot of product development and complexity in turning this complicated tech into something that’s easy to use.”

Some foreign tech giants’ AI work could end up helping Canadian startups grow. Pasupalak said less specialized engineers can now build products on top of basic infrastructure like conversational analysis tools and bot frameworks offered by Microsoft, Google and others. Amazon Web Services did something similar for app developers. 

“AI is democratized a lot more” than a decade ago, said Pasupalak, now an entrepreneur-in-residence at VC firm Inovia Capital. There’s a lot more talent.”