Ragavan has over 5 years of experience designing and deploying deep learning for enterprises from diverse industry sectors, ranging from agriculture to finance. Passionate about taking Canada’s AI industry to the next level, Ragavan designs and builds deep learning solutions that drive business value for enterprises around the world. Ragavan led DeepLearni.ng’s recent development of the world’s first deep learning system for retail banking at Scotiabank, which has already saved the bank millions of dollars.
Q: Please quickly introduce yourself and your organization? Why did you decide to join DeepLearni.ng?
A: My name is Ragavan. I did computer engineering for my undergraduate study, and also participated in a program called NEXT 36, which is an entrepreneurship and leadership program at the University of Toronto. During my undergraduate study, I started working with machine learning and deep learning. I grew up with Alex Krizhevksy, who is a research engineer at Google. When he released his ImageNet paper, I got really interested in deep learning, specifically his work with convolutional neural networks. I started my first company in 2012, and ever since then, have been applying deep learning in different areas. I’ve already seen deep learning have a large impact on many industries, and I really love the applied side of it.
What I noticed about Canadian entrepreneurs is that a lot of them tend to sell their companies to Americans, so we do not end up with really good companies in Canada. A lot of them would disappear or merged into big companies. I saw this as an opportunity to create something great and impactful in Canada. I am currently one of the co-founders of DeepLearni.ng and lead all the machine learning work over here. We saw that a lot of enterprise companies were really interested in deep learning and AI. A lot of people were trying to do things and not succeeding. What we decided to do is actually to go out to help these enterprise companies. We create and deploy pilot AI solutions and make sure it actually delivers values to the enterprise. That’s our main focus, we help large enterprise companies create AI.
Q: Why did DeepLearni.ng decide to sponsor aUToronto?
A: The main thing is that we love the community here, and we really would love to support it. The biggest challenge for autonomous driving is that it faces a lot of real-world challenges that you don’t typically see in the research setting. It would be something we have a lot of experience in, things that are not in the research paper and things you don’t think would affect the success of your model and the system you end up using. So, we saw this opportunity to really help there, and we want to support our community, and overall it is a cool project.
Q: What is DeepLearni.ng’s expertise related to artificial intelligence (AI) technologies?
A: In general, our focus is on deploying AI solutions in production for large enterprise companies. The challenge with that is you have to incorporate AI into old legacy systems. It is not like what you see in school. It’s about how do I make sure that my entire system is working for the users and for the community. You need to think about a lot of different challenges, and that makes the problem a lot harder to solve. There are a lot of externalities involved with these real-world projects. I will say that is our focus.
Q: What are your personal opinions on roles of engineering students to meet future society's challenges?
A: Engineering students have a really big burden in terms of how their responsibilities are. What we are starting to see now is something that traditional engineers have is professional practice in ethics. I think a lot of software engineers and computer engineers are kind of ignoring it. They think what they are making is just predictive software, and it is not going to be a bad thing, right? But we are seeing more and more of these situations and massive impacts from software engineers, like the huge scandal happened with Facebook. That should be something that we have to think about. Especially with AI, if we are creating biased systems, that would be incredibly impactful and dangerous. There is a heavy burden because now with the power of the tools we have, we not only can affect these things very adversely, but we also have the duty to make sure that they are used for goods.
Q: What’s one thing you wish somebody would’ve told you before going into the artificial intelligence field?
A: I think one of the things we tend to do as machine learning engineers and software developers in general is that we are kind of over-index on creating cool algorithms or models with the best accuracy. The important thing is to solve real problems, to help people, or to actually make a difference in the world. Instead of asking “how do I make the coolest thing or the best component?” always think about if it’s solving a real problem. That’s something we’re trying to focus on at DeepLearni.ng.
Q: What is DeepLearni.ng’s major goal in the next 5 years?
A: In the next 5 years, we want to grow our team massively and also want to create an amazing place in Canada that people can feel proud about things they have created--a place where people can make an impact.
Q: Any methods for students and our followers to stay connected with DeepLearni.ng?
A: We have a monthly newsletter which students can subscribe to on our website, which features a hand-picked selection of cool AI news, commentary and research. You can also find us on Twitter and LinkedIn. We also host a Meetup Lab for production-level machine learning, which you can learn more about here.
A couple of our other channels: