Meet Luis Serrano! Luis is an AI Scientist, entrepreneur and the former Head of Developer Relations at Cohere. Here, Luis shares how following your passions can be a zig zag path and incredibly fulfilling, his vision as a teacher and entrepreneur and what led him to build in Canada.
What's your story?
Oh, the long version? Let's see. Well, it's wild because I kind of emigrated to Canada three times. I first came to Canada for my undergraduate studies at the University of Waterloo. I wanted to study math because I got really into it in high school and was looking for more opportunities in the field. Back then, data science didn’t exist, so career paths for mathematicians weren’t very clear. But I loved Canada and the university, so it was a great decision.
I continued on an academic path, aiming to become a professor, which is a common goal for math students. I did a master’s at Waterloo, then moved to the US for a PhD at Michigan. After five years, I earned my PhD in math and did a postdoctoral fellowship at the University of Quebec in Montreal. That was my second time in Canada.
While doing my postdoc, I started getting interested in machine learning, which was becoming a big buzzword. I decided to change careers, left academia, and got a job at Google. Although I wanted to stay in Canada, I ended up moving to San Francisco to work at YouTube on their recommendation engine, which was a lot of fun.
I realized I really enjoyed teaching, so I switched to Udacity, an educational platform, where I helped create machine learning courses. This was around 2016, during the first boom of machine learning when it started to recognize things like cats in images, which was mind-blowing at the time. While at Udacity, I also started a YouTube channel - Serrano Academy - and wrote a book called "Grokking Machine Learning." The channel has grown significantly with almost 150,000 subscribers and millions of views.
After Udacity, I moved to Apple to teach at Apple University, which was a fantastic experience. I got to do internal consulting as well, learning from different teams about their projects. Eventually, a friend called me about a machine learning position in quantum computing, so I moved back to Canada to work in Quantum Artificial Intelligence. I was creating content on the side throughout all this, like YouTube videos and leading talks.
Most recently, I got into large language models before ChatGPT became popular. I joined Cohere in Toronto to build educational material for their LLM University. After ChatGPT’s boom, it became easier to explain my work. Three weeks ago, I quit my job at Cohere to focus entirely on my YouTube channel and build my own teaching and consulting business.
Teaching and learning new stuff have always been my anchor. The AI world changes so much, which makes it fun and exhilarating. That's my story in a nutshell.
That’s an amazing story! What are you most excited about with your YouTube channel, and what’s your vision for it?
I’ve always loved teaching because it brings out the giving side of me. I enjoy making learning approachable and fun with animations and quirky examples. I think abstraction is important, but it’s not the best way to learn. Simple, memorable examples work better for most people, including myself.
The fact that I can make a career out of this is a nice bonus. I love getting messages from people around the world who found my content helpful. It’s rewarding to bring education to those who might not have access to standard educational resources. So, the vision is to continue making engaging, educational content that’s accessible to everyone.
That’s fantastic. You’ve just gained a new subscriber! What predictions do you have for the future of large language models?
In the immediate future, I see multimodality as the next big thing. Models will handle text, sound, images, and video as input and output. Tool use is another development on the horizon. Models will be able to directly perform tasks, like sending emails or making web pages, instead of just providing instructions.
In the medium term, integrating AI with robotics is a likely step. Combining language models with motion and physical interaction could lead to exciting advancements.
In the long term, I get philosophical. As machines take over more logical and rational tasks, humans might need to focus on what makes us unique, like empathy, intuition, and other soft skills. These are areas where we can still outshine machines and should be embraced.
That’s insightful. Switching gears, what keeps pulling you back to Canada as the place you want to build in?
Canada is a wonderful country and has always been welcoming to me. It provided scholarships and funding for my research, which I’m very thankful for. I love the peacefulness and social benefits, like healthcare, which I didn’t have in the US.
Canada is also a hub for innovation, especially in the corridor from Montreal to Toronto to Waterloo. It’s home to leading scientists and researchers in AI and quantum computing. The vibrant research community and the welcoming nature of the country make it an attractive place to live and work.
What’s the most Canadian experience you’ve ever had, or a memorable moment here?
One memorable moment is when my grandma, in her 90s, got her Canadian citizenship. She was the oldest person in the room, and the judge came to take a picture with her. It was a special moment that highlighted how welcoming and supportive Canada can be, especially towards immigrants.
Σχόλια