Newly-Launched Canadian AI Startup Cohere Has Hit the Ground Running, Hiring a Multifaceted Group of Tech Talents in the Hope of Solving One of the Industry’s Most Vexing Issues: How To “Talk” to Computers… and Get Them To Talk Back
DALLAS, TX / ACCESSWIRE / May 11, 2021 / How's this for a conversation starter: Toronto-based Cohere, a startup founded by Ivan Zhang, Aiden Gomez, and Nick Frosst, has made it their mission to take AI to a new frontier where computers have a deeper understanding of how humans speak. Until now, the subtleties of speech, such as tone, sentiments, and semantics, have been beyond the scope of our mechanical creations. For example, you can ask Siri, Alexa, or HAL to "fix me a martini," only to receive "But Dave, it's not broken" in reply.
Laying the groundwork for what's certain to be a groundbreaking leap in bio-techno understanding was no easy task in itself, and it all comes down to people - no offense to AI.
Before co-founding Cohere, Aidan Gomez made waves in the AI industry by co-authoring a 2017 paper at Google that introduced a new architecture called the "transformer," which heightened the efficacy of natural language processing solutions that use algorithms to generate language.
Co-founder Nick Frosst worked with renowned British-Canadian cognitive psychologist and computer scientist Geoffrey Hinton, noteworthy for his pioneering work on artificial neural networks. Based at the University of Toronto, Professor Hinton is also a Google engineering fellow and has earned the sobriquet "the godfather of deep learning." Nick Frosst studied computer science and cognitive science at the University of Toronto and kicked off the Toronto Google Brain office with Professor Hinton. Over the next three years, Frosst performed intensive research on capsule networks, interpretability, and adversarial samples in machine learning.
Gomez and Frosst first met at Google Brain, where they found that by translating gains in natural language processing from research to practical applications, they could improve the working dynamic between people and computers. In 2019, Gomez and Frosst decided to leave Google and start Cohere along with Ivan Zhang, who had been working in a research lab with Gomez and had experience serving and scaling machine learning models.
Gomez', Frosst's, and Zhang's collective high profile in the burgeoning AI field has attracted the attention of a number of highly respected investors, including the aforementioned Geoffrey Hinton, Raquel Urtasun (Chief Scientist of Uber ATG), Ian Goodfellow (Director of Machine Learning in the Special Projects Group at Apple Inc.), Fei-Fei Li (Sequoia Capital Professor of Computer Science at Stanford University), and Sanja Fidler (Director of AI at NVIDIA).
Investment is only one side of the coin, pun intended. Conceptualizing and constructing a safe, secure, and practical framework for understanding between "natural" and artificial intelligence requires something more: the aggregated power of the industry's best and brightest minds. To achieve that benchmark, Cohere prioritized hiring and acquiring the very best people available. In a year, Cohere had expanded beyond its founders to a tightly-knit team of 25 employees, many (but not all) of whom came from top-ranked research laboratories and universities. It truly can be stated that in the age of "talent wars" among rival tech firms, Cohere has performed an extraordinary feat of talent acquisition.
"We're a small, diverse team with loads of experience in research and engineering," to quote Cohere's newly launched website. "We come from the best labs and schools in the world, and we've also dropped out of them or have never been to school at all! We welcome brilliant people from wherever they may come."
You may be wondering what a cutting-edge tech startup backed by numerous deep-pocketed investors is doing hiring people who have dropped out of school or, even more bizarrely, never even set foot inside an Ivory Tower. To get to the bottom of this conundrum, we took the opportunity to chat with Carol Chen, Cohere's youngest employee and an "uneducated" ML Ops engineer. Just eighteen years old, Chen was one of Cohere's earliest hires despite the fact that she never went to college.
We asked Chen how a high school graduate could gain enough knowledge to be able to work alongside degreed employees on some of the most important software the AI industry has ever created.
"When I interviewed (at Cohere)," explained Chen, "I was still skeptical of leaving my position at Shopify for a startup. I would be only their sixth employee. Even though the company was just getting started, Ivan convinced me it would be worth it. They were right - this software is truly incredible. Cohere is in a position to start bringing powerful AI to the world in a safe way, introducing it to us in an intuitive and non-invasive way."
You read that correctly. Carol Chen is an ex-Shopify software engineer. She started her engineering career even earlier at the age of fifteen, garnering countless high-profile hackathon victories, winning the 2017 Google Code-in, publication in Shopify's engineering blog that was cited by Oracle, and gaining practical experience working on compilers. Even with all that (and more) under her belt, did she have any problems adjusting to her new position at Cohere?
"I work on scaling the serving of large language models," explains Chen. "That involves not only writing code to manage our servers and accelerator hardware but also making the models faster. I think many startups try to build software as quickly as possible and then make it fast and scalable later. At Cohere, however, we believe that the software needs to be excellent in the first place."
There's much more to job satisfaction than just being able to do the job, of course, and Chen says that working at Cohere has been fun thanks to her coworkers. "They value experience and knowledge over formal education, and that's why I've been treated very well from the beginning. Also, they are just great people to work with, which makes work more enjoyable and improves the quality of the work I produce. My coworkers at Cohere are talented but also fun, kind, and all-around lovely." The close-knit and collegial work environment at Cohere has provided extra benefits for Chen, as she's been able to increase her ML (machine learning) knowledge simply by "learning on the job."
After working at Shopify (as of 2021, the largest publicly traded Canadian company by market capitalization) and now at Cohere, Chen sees some differences between being employed by a very large publicly-traded company and by a small but growing startup.
"Ideally nothing, and at Shopify, it certainly varied at different teams. One thing that is similar that I loved is that at Shopify interns are treated well and work on real projects, and it was not uncommon for people to not even know that someone is an intern. However, I feel the essence is very similar (compliments to Shopify), as I always felt like the team was moving cohesively and were not subjected to any politics. The main differences are in knowing what's happening in the company. At Cohere, I know everyone and everything going on; at Shopify, I couldn't even begin to know everyone or everything that was being worked on."
Kudos to Cohere - they might be "only" a startup with all the pressures that engenders, but they're still able to provide a work environment as encouraging as the one at Canada's most attractive tech employer, Shopify. Could this be the secret of attracting and keeping top-notch talent when you're the new kid on the high-tech block?
Speaking of new kids, we asked Chen how she got started as an engineer in the first place, back in her early teens. Her answer was as unexpected as it was revelatory: "I just liked doing things. Unlike practicing surgery or defending cases, there was nothing to stop me from building things with software and getting people to use them. I built a lot of small projects at hackathons, which led to my first job. The experience I gained there allowed me to get my next job, and over time, that translated to knowledge and skills that led to my position at Cohere."
Chen's career path has been atypical to say the least: she turned down acceptance offers from the top two Canada's computer science programs (University of Toronto and Waterloo) in favor of entering the workforce on her own terms and based on her proven abilities. Interestingly, now that she's established herself at Cohere, she has students from Waterloo University working as her interns.
As can be expected, Chen gets a lot of questions from high school and college students about the value of getting a university degree. Her answer is that it's definitely needed depending on the field, but in some fields, it just doesn't count as much as it used to. "It certainly isn't needed in software engineering jobs," she states, "and though it doesn't come without challenges (some people aren't as progressive yet) but four years and six figures of schooling costs give self-taught ‘students' like me a big head start. Take it from me, even for computer science research jobs in the industry; there's a way in without a degree."
That unorthodox path has led Chen to Cohere, where her and her coworkers' efforts to build machines that can understand the world and be safely used by everyone continues. Shaped by their varied skills and bolstered by a uniquely catalytic combination of age, experience, education and homegrown tech-savvy, it is a mission that is likely to succeed. Being a pioneering company in an esoteric field, Cohere realizes that acquiring the best talent means looking for minds that think outside of the computer case. The letters after their names are of no consequence if the stuff between their ears can't cut the mustard.
This creative and unconventional method of talent acquisition bodes well for Cohere. It's no secret the tech industry has dealt with issues such as exclusivity, ageism, and misogyny - issues that still linger and are difficult to resolve considering they emanate from the executive suite. By focusing on "what" instead of "who," Cohere can potentially build a strong employee brand that will pay dividends through their shared successes.
The benefits aren't exclusive to Cohere, either: provided with the opportunity to excel based purely on their skills and without being pre-judged, Cohere's employees are free to apply their knowledge to push AI - and humanity's understanding of it - to new heights that will vastly benefit society as a whole.
CONTACT:
Cohere
salevenstein@gmail.com
SOURCE: Cohere
View source version on accesswire.com:
https://www.accesswire.com/646638/Helping-Humans-Converse-With-Computers
