Tips for Getting Started Working in AI — Paige Dickie, Head of AI at Layer 6
We had the pleasure of chatting with Paige Dickie, Head of AI at Layer 6 on the Women in AI Podcast. Paige began her career in management consulting, working across topics from data strategy to digital transformation. More recently, Paige worked at the Vector Institute for Artificial Intelligence, where she led initiatives with Canada’s largest financial institutions, consulting companies, regulators, and Government agencies, before joining Layer 6 earlier this year. At Layer 6 Paige is globally responsible for the end to end workflow and life cycle of all use cases across the bank.
🎧 Listen to the podcast here or watch the video below.
Topics explored include:
- The Exciting Potential for AI in the Future
- Canada’s Contribution to AI, the Institutes and Its Influential Thinkers
- A Day in the Life for Paige at Layer 6
- The Integration Between Layer 6 and the Bank
- Advice for Beginners Entering Their AI Career
- AI for Social Good & Ethics, and What Social Benefits Can It Enable?
- Tips for Women in AI and Role Models
Paige, it’s fantastic to have you join us today. It’d be great if you could just tell us a bit about your background.
Firstly, thank you so much for having me on, I’ve been very excited about doing this. It’s kind of a unique opportunity that I haven’t done before. A little bit about my background, after completing a number of advanced engineering degrees and biomedical mechanical engineering, I decided to make a little bit of a right turn into management consulting. I dabbled with a start-up during my master’s, and that really just made it painstakingly apparent that I needed a crash course in all things business. So I joined McKinsey where I spent a good chunk of my career consulting companies, fortune 100 companies on things like data strategies and digital transformations or even stand up innovation centres.
I guess this will resonate, probably with anyone who’s done management consulting, but over time, the airports, the airport food, the flights, the hotel, it all wears off, and so I decided to transition into industry. I spent a couple years at the vector institute before I was approached for an opportunity at TDs AI lab and i’ve been here ever since.
So kind of jumping straight in there with quite a big question. What do you find exciting within AI currently and also looking a bit further ahead for its potential for the future?
Great question. You’re probably looking to hear a little bit about some of the things that I think are the leading edge applications coming down the pipeline that we expect to hit the industry soon. I could easily talk about Graham Taylor from the University of Guelph, he’s using AI technology to make the agricultural field more efficient food systems. He’s developing smart insect traps that are using image recognition to identify pests, and it negates the need to actually have to send those samples off to labs, reducing overall cycle times in the industry. More I could highlight, Frank Rudzicz’s company Winterlight, who’s using speech recognition, they provide an accurate objective measure of one’s cognitive health.
Basically users describe an image that they see on a tablet and based on you know, hundreds of variations in pitch, vocabulary, and grammar, these things combined to predict an accurate objective measure of things like depression or the early onset of Alzheimer’s disease. So I think those two applications are absolutely fascinating. But if I were to say what the front running thing that I think is exciting about the future of AI is, I think in many ways, it’s actually kind of Canadian, if that makes any sense.
So what do you mean exactly by its Canadian?
Let me back that up a little bit. Firstly, I do not mean to discount the incredible contributions that folks have made all around the world, they’re absolutely brilliant people contributing to the field globally. It’s just that some of the most prominent and influential thinkers within the field of AI are actually Canadian. I can give you some examples, there’s Rich Sutton who’s pioneering Reinforcement Learning at AMI in Edmonton, which is a one of Canada’s three AI Institutes. You have Geoffrey Hinton, many refer to him as the Godfather of AI, and him and Yoshua Bengio, Yoshua Bengio heads up MILA in Montreal, another Canadian AI Institute.
They actually just recently won the Turing Award for their contributions to the field of AI. So we have some really incredible theoretical research talent here in Canada. Canada was the first country in the world to come out with a national AI strategy, and the reason they did this is that a lot of our AI talent would actually end up leaving our jurisdiction to go to the US and become the chief AI Officer of Facebook and Google.
So the government said we’re experiencing a lot of brain drain, how can we reverse this? I kind of touched on two of the three AI institutes. There’s also Vector in Toronto, which I mentioned, I spent a couple of years at. The intent of these institutes was to increase the production of Machine Learning talent in the country so that labs wouldn’t pick them off from abroad. Tech companies would actually relocate and open up shops in our economy and in the past decade we’ve seen countless labs open up in the area which just gives these researchers a breadth of more exciting opportunities. I guess it seems like Canada has that AI it factor and I’m excited to see how that evolves.
It’s something that we’ve actually seen through our events as well. We held our first Deep Learning summit in Montreal a few years ago and we’ve kind of alternated with Toronto since we’ve seen that grow year on year. So you’re doing something right I think. Taking a bit of a step back, what does your day to day look like for you in your current role as head of AI Business Management at Layer 6 AI TD?
A day in the life of me, I spend a good chunk of my time meeting with lines of business across the bank to really get a better understanding of what their challenges are and whether or not AI can play a role and help them out those challenges. And to do this I lean a lot on our Layer 6 Machine Learning scientists, on their expertise and our engineers to explore the art of the possible and also really to keep me grounded. In reality, the dream is big and they bring me back to reality.
Within the bank our TD partners, we work time, or I guess I spend a lot of time with our control partners, really keeping layer six in line and also with TDs model validation pioneers, they’re really leading-edge thinkers who hold us to some of the highest standards in the Western Hemisphere. It’s their job to validate our models, integrity, and ultimately to safeguard our financial sector and they’ve done an incredible job in keeping up with the pace of innovation. I guess the other team that I spent a tonne of time with is our enterprise machine learning team. We truly wouldn’t have had any impact without these folk, so a special shout out to Matt, Zayn, Nuresh, Chris and Gwendolyn, if you’re listening, you guys move mountains for us on a daily basis. So, thank you guys for enabling us to be successful.
Well, it sounds like a lot of stakeholder management there.
Just a little.
So Layer 6 was only acquired a couple of years ago. And how has, since then integration with the rest of the bank been playing out for you and your teams?
With all acquisitions, we’re a small company in a massive bank, so there’s really lots to learn and a long way to go before we can truly claim we’ve transformed ourselves, TD Bank into the world’s leading AI-powered bank. And to be clear, that is the goal. We’re a bunch of stubborn folks who won’t stop short and the ambition on this team is stifling. We have the incredible talent inside TD and inside Layer 6, and I honestly wouldn’t put my money on another horse.
I guess speaking a bit to the listeners that are maybe at the beginning of their career or looking to advance, what would you say the best way to do that is in industry, and also what path did you take to get to where you are now?
So there are so many different paths in the industry to get into AI, that’s a very open-ended question. But I can certainly share my own experience, I left McKinsey to join TDs digital channels doing digital strategy and innovation. And under the innovation portion of my mandate, I was exploring all these emerging technologies like blockchain and IoT, and AI. I guess from where I was sitting AI appeared to be the heaviest hitter and the ethical considerations at the time just really fascinated me and so I knew I wanted to pivot to build a foundation in the field.
I actually proactively sought out opportunities which basically means I started putting my hand up for extra work, those stretch assignments, and so I would take little pieces of AI assignments all across the bank and I also started asking our HR group how I could progress and they were incredibly helpful. I owe a lot to them. I guess this is one of those ask and you shall receive principles. It’s either that or people in the bank got so annoyed with me constantly asking to be involved in AI that when a secondment opportunity came up with the Vector Institute, they knew it was the only way to get me off their backs.
That’s a good way to do it.
So jokes aside, I owe a huge thanks to TD’s head of digital channels, Rizwan Khalfan & their head of HR at the time, Daniel Decotis. They truly took the time to understand what my passions were and to build a bridge for me to get there. So if I were to give everyone a pro tip, surround yourself with people who will invest not just in your company’s mission and vision, but in your own.
You touched a bit on ethics there as well, it’s something I want to dig a bit deeper on. A topic that regularly comes up during our podcasts is looking at AI for social good and more about the positive impact we can have. So I wondered if you could just tell us a bit more about your particular interest in that and what you’re working on in the AI for good space?
Let me start by saying AI is an incredibly powerful tool and so for all the marvel fans out there, we all know that with great power comes great responsibility. The idea of applying these technologies to do some good for the world rather than put another dollar in an investment banker’s pocket. So let me share some of that, I have a couple of examples that come to mind.
One of them started not too long ago when I discovered the prevalence, the unfortunate prevalence of human trafficking in our country. I think we can all agree that one case is just one case too many and as many of you know, these heinous crimes generate revenue and these criminals often tried to launder these funds through the banks and the banks have an obligation to identify and stop these transactions from happening. Which to me means it’s actually an area where Canadian banks can come together alongside law enforcement and AI experts to stop these crimes in their tracks.
One of the reasons I actually think Canada is so well positioned. Canada has five big banks and those big five banks, bank about 90% of the Canadian economy and not only that they’re actually all headquartered in Toronto. So it’s, it’s just really easy for five people from five banks to get together and collaborate on something like this. Versus let’s compare ourselves to the US and their five big banks, their bank is roughly 35% and you have 6000 some odd other players dispersed across the country, which just makes these kinds of joint collaboration efforts much more challenging.
It’s incredible how banks can leverage that technology to fight those kinds of crimes, really. But what about helping people with kind of more, I guess, basic banking needs? Are there any social benefits that AI can enable in that space as well?
Another area of interest centres around helping the financially vulnerable populations. Around half of the Canadian population is living paycheck to paycheck. And even a small incidence, a broken furnace or a fender bender can send the most well-intentioned people who are planning well into a financial death spiral. And I guess another contributing factor to this as our economy is shifting towards a flexible on-demand workforce, perpetuating the already existing income volatility issues for again over half of Canadians.
So like a mistimed inbound check in an outbound check means they could get hit with things like late fees or, you know if you miss a payment on your credit card, payment of interest fees or just this compounds our situation further and other people who need it, the least in our economy. And so similar to how the flexible workforce economy is using AI to predict demand and vary workforce accordingly, AI can play a role here and helping predict someone’s cash flow and you know what their account balance are they going to go into that insufficient fees and get that hit and if we can predict that in advance and give people notice, we’re giving them a better line of sight and really control into their situation so that they can improve their financial outcomes.
That’s fantastic. And I wanted just to touch on, I guess a topic that’s pretty central to this podcast. So as yourself you’re a leading female working in AI at the moment. What piece of advice would you give to your younger self when you are joining the field and first starting out?
Great question. Jump in with two feet. I think a lot of times we shy away from things or we’re worried about getting things wrong and these hold us back. AI is going to be a foundational knowledge block that’s going to underpin so many future jobs and having a firm understanding of the basic concepts and capabilities and, and even hurdles and limitations. It’s just from my point of view, it’s a no-risk manoeuvre for a very large portion of people. And if you really want to transition, Toronto, like I said has some of the world’s top global AI talent. You know you can check out the amazing programmes that are offered by the Vector Institute or MILA in Montreal and AMI and Edmonton.
There’s a tonne of online courses offered through UoT. And again, the godfather of AI, Geoffrey Hinton came from there. There are incredible programmes are at McGill. There’s a lot of great courses and programmes that are offered online through these institutes and our educational systems. And of course, they’re great podcasts like this, which I think is another option for people to just get their feet wet in the field.
Exactly. And what about specific areas of focus? So if the foundational skills that are important for a career in AI now?
So I guess for some of the younger folks who maybe want to think about picking their courses, I think having a good understanding in statistics, math, and computer science, in general, will set you up for a successful career in the field.
Great, a pretty good place to start. And what about yourself? Do you have any female role models in the fields that you would if you had to pick to share with our listeners?
One of the forefront people in my mind is Marcia. She’s in the public health space and from vector and she’s helping to make sense of messy clinical data. So in the current COVID economy, I think her role is going to be critical, and I’m excited to see what comes out of this next but just to give you some background, she’s designed a suite of machine learning methods that can predict how patients will fare during their hospital state. Her algorithms can accurately predict things like how long the patient will stay in the hospital or are they likely to pass away while they’re there or even if they’ll need interventions like blood transfusions or ventilators.
And actually, she was named MIT Tech Reviews List. She was on the list of MIT Tech Reviews top 35 innovators under 35, which I just think it’s an incredible accomplishment. There’s Patricia Thane, she’s doing some incredible work in the AI security space with her start-up private AI. She’s also an incredibly savvy, well-rounded founder who I think just has a really bright future ahead of her. And maybe I’ll bring up Raquel Urtasun. She’s actually the head of Uber ATG in Toronto, which, for those who don’t know, is the Innovation Lab dedicated to research on self-driving vehicles.
She’s also an assistant professor at the University of Toronto in Computer Science. She’s a Canadian Research Chair in Machine Learning and computer vision. And she’s a co-founder of the Vector Institute and Uber recently announced that they’re going to invest actually wasn’t recent, I think it was maybe two years ago now, but they’re going to invest $200 million and employ more than 500 people over the next five years in a software engineering hub that’s actually going to be headed up by Raquel Urtasun. They are incredibly talented females, they’re powerhouses and they’re pushing the boundary of the status quo. And so I just I love to see their successes.
It’s impressive news. And Raquel has spoken at our events previously and always has, well just is really inundated with questions from the audience. We can never have long enough Q&A. So, so some great, great names there. And what would you say needs to be done on a bigger level, or higher level, to encourage more women to work in AI in the future?
I think there’s many things that can and should maybe be done and to name a few. I think one of the helpful things for getting women into a field is to have female role models. Seeing women doing those roles, it helps people picture themselves in the role and it really avoids them developing self-limiting beliefs. I think we impose these restrictions on ourselves, and seeing somebody else in the field just immediately removes that boundary that we imposed on ourselves and kind of becomes less intimidating. Another area is taking the other lens and looking at how companies can attract more females is your marketing tactics and strategies.
Maybe you aren’t proactively recruiting in a gender-neutral fashion, or maybe even pro-female tactics to just increase that funnel pipeline, given it’s drier than the male pipeline. And then I think there’s a lot that we can do for controlling for systemic gender-based discrimination. We do this in all of the AI models we build and so I think it’s if we can do it in an AI model, we can do it in an institution. And the odd thing is I always hear these debates coming up about, you know, bias and the potential for discrimination and models and by all means, that is true, these models can go awry, but I think the important thing to take away here is that they are built on historical data.
So these models aren’t biased inherently, they are reflecting the biases in our historical data, which is really just holding up a mirror to our historical society and saying, our history is biased. And I think that’s bringing to light, really, really important issues that we need to address. I actually think AI has a role to play in, in stopping this kind of gender-based discrimination or any kind of discrimination for that matter.
Well said. Thank you so much, Paige, for taking the time to chat with us today. I think some of the advice that you’ve given to our listeners will be really, really kind of grateful for those of our listeners that are just working in the field kind of at the beginning of their career and starting out. And also, it’s been great to hear a bit more about the work that you’re doing at TD Bank and Layer 6. So thank you again, and I hope to see you at one of our events at some point soon.
I would love to be involved. Thank you so much for having me. This is wonderful. I hope that something I said was useful to someone, and until next time.