Matt Zeiler: How an EngSci taught computers how to learn
By Michael Klassen, 2009 fellow graduate of EngSci.
I recently stumbled across an old EngSci 0T9 class photo and I was struck by the remarkable dreadlocks sported by my friend from rural Manitoba, Matt Zeiler. As classmate and friend, Matt was on the quiet side: he liked to party, but he spent much of his time running, playing basketball and studying intensely. His academic excellence and EngSci ranking brought with it some notoriety, but that all pales compared to his recent success as an entrepreneur. Matt is founder and CEO of Clarifai, a company than in in simple terms provides a “solution that allows computers to see”. Using a set of computer algorithms called neural networks, Clarifai is able to process and understand visual data with unprecedented speed and accuracy. If you are curious to learn more, you can actually test the algorithm on their website for free, and see how well it can pinpoint both objects and more abstract themes in your images.
So how did the quiet, dreadlocked EngSci suddenly catapult to the forefront of the artificial intelligence scene, giving interviews on Bloomberg and being featured on a monthly basis by magazines like Wired? Let’s start from the beginning.
Matt actually started his university education closer to home, completing one year of Pre-Med at U of Manitoba before making what he deems to be the most significant life decision to date: switching to Engineering Science to ‘start over’ at U of T. He credits the two-year foundation of broad and deep overload courses as teaching him how to adapt and learn quickly in the face of large volumes of difficult information. Matt remembers the value of “starting from scratch to work through the entire process of engineering design” in the legendary AER201 Engineering Design course, as well as “having to put things in perspective in the real world” during CIV102.
A crucial inflection point in Matt’s EngSci career was his introduction to an eerily lifelike video of a fire produced by computer algorithms. This influenced Matt to change allegiance from his plan to study Nanoengineering to instead go into the Computer Option and tailor his curriculum to machine learning and neural networks. Matt credits the clout and influence of EngSci in getting Geoff Hinton, known as the godfather of neural networks, to supervise his 4th year thesis.
In second year, while building a robot and keeping up top grades, Matt started his own business, ReviewMate, to market and sell ‘best in class’ review booklets tailored to specific EngSci courses. This proved to be a living laboratory for entrepreneurship, from product development to marketing, advertising and sales. He even roped his future wife into helping put together the very first flyers for the business! ReviewMate navigated an evolution to a peer-production model and a shift to online sales, taking the pressure off Matt and drastically expanding the market to other faculties and eventually other universities.
Matt carried the momentum of his undergrad directly into a PhD in Computer Science at NYU, where again he was able to work with world experts on computer vision (Rob Fergus) and neural networks (Yann LeCun). Matt immersed himself in the world of machine learning and artificial intelligence, and worked again with Hinton and the extremely influential engineer Jeff Dean during two summer internships at Google in Mountain View, gaining exposure to more practical applications of neural networks.
By the time he finished his PhD, Clarifai was up and running out of his apartment. As of last week, the company has since grown to 12 in its own Manhattan office, and an impressive bench filled with industry experts and recent PhD grads. With the recent announcement of $10million in Series A financing, Clarifai is starting to build out its business and sales team to compliment the unbelievable depth of engineering and computer science talent it already has to offer.
These days, Matt is riding the waves of being able to hire ever more “A players” as he describes them, and representing the company at conferences and in interviews daily. Of note to EngScis: Clarifai is currently hiring significant numbers of new staff, and is open to finding ways to create internship and even PEY opportunities.
My own path couldn’t have differed more than Matt’s since we met in 1st year EngSci, and it is remarkable to feel the motivation and peer challenge of a classmate making such big strides at such a young age. In that way, the multidisciplinary ‘machine learning’ ways of EngSci have impacted us both.