If AI takes care of the difficult work, how do you keep your mind trained? Think of astronauts exercising in space: without the usual pull of gravity, they need to make a deliberate effort to keep their muscles strong. Learning with AI raises a similar question. When a tool can produce answers in seconds, you still need to practice the thinking that helps you understand and question them.
For Dr. Lino Coria Mendoza, that sometimes means asking students to close their laptops, put their phones away and work with pen and paper. But on the other hand, he teaches students to build AI tools that make their lives easier. What matters is knowing when a tool helps you learn and when it lets you skip the work needed to understand.
Dr. Coria Mendoza is a teaching professor at Northeastern University’s Vancouver campus and director of its Master of Science in Artificial Intelligence program. His more than 20 years across industry and academia include almost a decade at Vancouver startups, developing tools for YouTube creators, custom footwear and robots that monitor greenhouse crops. Teaching was always his first ambition, and the talks and workshops he gave while working in industry helped bring him to Northeastern’s attention.
Today, he teaches students who want the kinds of jobs he once held. In this Studyportals Campus Conversations interview, he explains how practical experience shapes his teaching, why students need to justify their decisions, and what they learn from working with people who think differently. If you’re considering an AI degree, he offers advice on choosing a program and developing the technical knowledge, judgment and communication skills to make good use of AI.
(Photo courtesy of Dr. Lino Coria Mendoza)
From industry experience to university teaching
Studyportals: You spent around ten years in industry before joining Northeastern, working on machine learning and computer vision solutions in Vancouver-based startups.
Dr. Lino Coria Mendoza: I worked at three startups, each in a very different field.
At one of them, we were building tools for YouTubers. I’m talking about 2012 or 2013. People making videos for YouTube needed extra tools. I was creating a tool, for instance, that would analyze their videos, identify the most relevant and interesting frames, and offer them as thumbnails. Because when you are watching YouTube and have a bunch of thumbnails, you click on the one that looks the most appealing to you.
Then I moved on to a startup that was making custom footwear. Footwear made just for you. We built a computer vision application. People would take pictures of their feet with their mobile phones. We would identify the shape of the foot, the length and the width, and create insoles that were the right fit for each foot.
At the last startup I worked at, we were building robots for greenhouses. These robots would go through each row of a greenhouse: tomatoes, cucumbers, peppers. I was in charge of the computer vision aspect. We would count all the tomatoes, let’s say, and classify them by ripeness. We would also include information about diseases and things like that. Then we would make projections: “You will get this many tomatoes in one week, this many in two weeks. But these areas of your greenhouse are challenging, so you’d better take care of them. Otherwise, this is going to happen in a week, in two weeks, in three weeks.”
Studyportals: These three startups belong to different industries. What did they have in common that allowed you to be part of these initiatives?
Dr. Lino Coria Mendoza: All three involved computer vision. Getting the right thumbnail meant teaching computers to identify what made a good thumbnail, right? Sometimes you used rules, and sometimes machine learning. Instead of writing rules specifically, you give examples: “This is good. This is good. This is not good.”
With the footwear, it was taking pictures and identifying the feet in the picture, their shape, length and width.
Finally, with the greenhouses, it was not just identifying all the fruit. This made people’s lives better. It’s impossible for a human to count all the tomatoes in a row.
Studyportals: At what point did you say, “I want to teach others what I do here”? What brought you back to academia after those years in industry?
Dr. Lino Coria Mendoza: Academia was always plan A, and industry was plan B. When I finished my PhD, I moved back to Mexico and was a professor at a university there. Then, for personal reasons, I decided to return to Canada. For a little while, I came back to the University of British Columbia, where I did my PhD.
But the most interesting jobs I found in Metro Vancouver, where I lived, were industry jobs. A lot of my friends from my PhD years were in industry, so I joined them. I was having fun, but I always had this idea of returning to academia.
I was a little concerned that I was getting older and was still a research engineer or a computer vision engineer. I thought: “I need to move up.” Even though I was doing some team lead work, on paper I was still a computer vision engineer.
So I thought: “I need to find a way to stand out. What do I have that most engineers don’t have?” And I realized I have great communication skills. I can explain complicated things to other people.
I started giving talks. If you invited me to teach a workshop, I would be there. That is how Northeastern learned about me and reached out.
Studyportals: How did that industry experience shape the way you teach and the way you are as a professor?
Dr. Lino Coria Mendoza: Oh, it made me a better academic. I was a professor before I joined industry, and then I was a professor after a decade of working in industry.
First of all, I am teaching master’s students who want to do the jobs I had. So I know exactly what tools they need and what tools they don’t.
I think there is a great advantage in going deep. If you really go deep into things, you get a better understanding. You can not only replicate something but go beyond it. But I also understand that in industry, many things are fast-paced. There’s a deadline; it’s in two weeks, and you need to deliver something.
I am teaching master’s students who want to do the jobs I had. So I know exactly what tools they need and what tools they don’t.
So there is all of that: how to solve problems, and also how to ask for help. When you’re stuck, you have your peers, and you need to ask them for help. They are busy doing their own work, but you have to find a way of getting something from them and continuing your work.
(Photo courtesy of Dr. Lino Coria Mendoza)
Learning to build AI, not just use it
Studyportals: How do you teach students to switch from a user mindset to a developer mindset?
Dr. Lino Coria Mendoza: I guess it’s like the difference between going to the dentist and being the dentist. We’re all worried about our teeth, but some people can actually fix things or help you have healthy teeth.
I think our students need to be really strong programmers, even if AI can write their code. AI tools can write better code than I ever could. But what matters is whether students understand what the code does and why it works that way.
They have to think: “How am I going to design this very complex system so that it is scalable? If something changes, can we remove it and replace it with something else? If more people use it, how can we make sure it doesn’t crash and can keep growing?”
Our students need to think about all these possibilities. They also need to understand how AI works on the inside so they can optimize their code to work well with it.
I think our students need to be really strong programmers, even if AI can write their code. AI tools can write better code than I ever could. But what matters is whether students understand what the code does and why it works that wayCode that performs the same task can run in ten seconds or ten hours, depending on how it is written. Here we go into math, right? Math allows us to understand how AI works and ask: “How can I write my code in a way that is most efficient?”
Studyportals: When it comes to the student mindset, what do they need to understand to create code that solves a problem in ten seconds rather than ten hours? How do you train that?
Dr. Lino Coria Mendoza: First of all, it’s knowledge. You need to understand how things work. We make sure students go through that. We explain why one tool is better than another. We look under the hood and identify things.
We need to be good at math and make sure we measure whether something has improved. It’s not just speed, right? You need to look at accuracy and a bunch of other parameters.
You also need to avoid fooling yourself. Sometimes you train or test with a small dataset, and you need to make sure your claim is meaningful. Do you need a bigger dataset? These are things students need to know.
Then they need to get experience. It’s practice: “This worked. This didn’t work.” Some things don’t come from somebody teaching you. They come from being there, trying things out and finding that something works or doesn’t.
Finally, something that is very important to us, and I hope to everybody, is the ethics part of AI. Who is going to be affected by this thing I’m building? How do I make sure I understand what those people need, whether they are affected positively or negatively?
Something that is very important to us, and I hope to everybody, is the ethics part of AI. Who is going to be affected by this thing I’m building? How do I make sure I understand what those people need, whether they are affected positively or negatively?
Talking to people is important too. I’m not saying we do it in every course, but we make sure there are moments during the coursework when students have to reflect on it and, ideally, talk to people.
Studyportals: All of this also needs to be reflected in the curriculum, and this industry is probably one of the most agile out there. How do you keep the curriculum relevant when technology develops so quickly?
Dr. Lino Coria Mendoza: The first thing I’m going to say is that most of the things I teach, I did not learn at university. I finished a long time ago, when these tools did not exist. But I was able to learn them because I had a solid background.
My math was solid, my programming skills were solid, and my understanding of machine learning principles was solid. So now, when you’re reading something, you can connect it with your previous knowledge. You’re learning about a new idea or tool, and you can immediately start connecting it with things.
Whatever we are teaching right now, I am aware that in five years there will be something new. But hopefully our students have the background that allows them to adapt.
Keeping the program fresh can sometimes be challenging because you cannot just switch courses. You cannot just say: “We’re going to throw this away.” There is regulation, you need to apply, and it gets reviewed by people outside the university.
But as part of our electives, we have special topics courses. These have a very generic title, and they give professors an opportunity to teach something for the first time.
Whatever we are teaching right now, I am aware that in five years there will be something new. But hopefully our students have the background that allows them to adapt.
I’ll give you an example. A colleague had an idea for a generative AI course when there wasn’t one available. He designed it and taught it first as a special topics course. It was well received, people liked it, and then it became a course.
Even if it’s not yet an official elective because of the process, when it is offered and our students want to take it, I review it. If I know it’s relevant, our students can take it as an elective. These special topics courses allow us to bring new knowledge into the university.
(Photo courtesy of Dr. Lino Coria Mendoza)
Using AI while developing your own judgment
Studyportals:You’ve also been exploring how generative AI can support student learning. Can you share some examples from your teaching and what you’ve observed yourself?
Dr. Lino Coria Mendoza: Sure. This is me as an educator rather than the director of the master’s in AI. I teach my students and look at when generative AI can be useful and when it cannot. How can we take advantage of it without being affected negatively?
I’m currently teaching a computer vision class, and it lasts two hours. For one hour, I’m going to ask my students to close their laptops and put their phones away. I’m going to give them a workbook, and they are going to do things by hand.
I want them to understand a concept very clearly and do some basic math. It’s not complicated, but it forces them to understand the process: from here, we go there.
If I’m just teaching and they have their laptops open, there’s no friction. They can get some numbers without ever really going through the process.
Learning has to be a bit of a struggle. If you don’t struggle, you’re not learning. It has to be a controlled struggle, right? You find something and you’re not sure how it works. You read a little more, try again, ask someone and go back to it. That’s how learning works.
Learning has to be a bit of a struggle. If you don’t struggle, you’re not learning.
But in other classes, this approach isn’t necessary, and AI tools are used to give students quick feedback on how they’re doing.
For instance, when I was teaching a research capstone course, students would write a paper. We created a tool with ChatGPT, back in around 2023 or 2024, where they could upload their paper and get feedback on what was strong and what was weak.
This was not to replace me or my teaching assistant reviewing the paper. Students would still have humans look at it and offer feedback. But sometimes having feedback from two people plus the AI tool helps you understand what needs improvement and what is working.
Studyportals:In projects that involve teamwork, students have AI tools to work with, but they also need their peers’ feedback. How do teamwork and feedback help students think critically about what they build with AI?
Dr. Lino Coria Mendoza: If you ask me what one of the most important tools any human being needs in life is, it’s critical thinking.
When you’re watching TV or going through your feed, critical thinking allows you to ask: “Is this true? What is the evidence for this? This person is saying all these things, but what’s the source?”
I get messages from relatives claiming things about politics, and I’m like: “What is the source?” Critical thinking goes beyond work and education and into everyday life.
We need to make sure our students can justify their decisions. They should decide to build something because they reviewed a bunch of options and understand the needs of the users or customers, not because an AI tool told them: “This is how it has to look.”
They need to understand the time constraints and the budget. When you go into industry, you have to be fully aware of these things. You cannot just build whatever. There are time constraints, a budget, people who want things a certain way, and decisions that have already been made that you cannot change.
We need to make sure our students can justify their decisions. They should decide to build something because they reviewed a bunch of options and understand the needs of the users or customers, not because an AI tool told them: “This is how it has to look.”
I want students to go through this and have conversations among themselves. These chatbots, like Claude or ChatGPT, could be your assistants instead of you being their assistant, right? Make sure they are supporting you and your thinking.
As an educator, I want to make sure my students can justify every decision they made in whatever final project they submit.
Preparing for work and choosing your program
Studyportals:What is different now for a student who graduates and wants to enter the job market? How has getting that first opportunity changed from back when you started?
Dr. Lino Coria Mendoza: I have students who are getting co-ops, and then I want them to be my teaching assistants, but they cannot because their co-op has been extended.
Or they’re so busy focusing on job opportunities that they don’t want to be my teaching assistants. That’s because they have great career opportunities.
What do these successful students have in common? I would say they always go above and beyond.
You ask them to build something that can do this, and they do it, but then they also make it do something else. They go beyond what was asked of them. They’re enthusiastic about it and explain why they are making it better than you originally imagined.
They also have great communication skills. They not only build something that works amazingly, but they can explain why it is amazing and how they built it.
What do these successful students have in common? I would say they always go above and beyond.
They can explain it to very technical people. But if they go to a hackathon and some of the people reviewing the projects are not experts, they can also explain it in a way that makes sense to non-experts.
Teamwork is a challenge, I would say, for most of humanity. But for students, it can be very challenging, especially because they are not necessarily a cohort that has been studying together.
They take electives, and there are different levels of expertise. Sometimes a first-year master’s student and someone in their last semester take the same course.
That is challenging, but they learn to navigate the politics and dynamics of working with teammates, even if they don’t love it.
Studyportals:You’ve studied in both Mexico and Canada. How does international study experience fit into this?
Dr. Lino Coria Mendoza: I think leaving home is a bit painful for the first six months, at least it was for me. It doesn’t have to mean moving to a new country. You can move to a different province, or to a different state.
Because Canada is so diverse, there are so many other international students. It felt like I got a tour of the world just by going to one city. The people in my lab, my department and my house were from different parts of the world, studying different things.
Because Canada is so diverse, there are so many other international students. It felt like I got a tour of the world just by going to one city.
I got a formal education, my master’s degree, when I first came to Canada, but I also learned so much beyond that. Ideally, this makes you more empathetic and more aware of differences in the world.
When people talk about self-driving cars in San Francisco, I’m like: “Yes, why don’t you bring them to Mexico and see how they work?” People cross the street wherever, and it’s such a messy driving situation. You can end up building something based on what you know without thinking about everybody else.
I think this is very important: students listening to people who think differently from them. It’s important to be exposed to different points of view, even when those views are hurtful to you. You want to understand why people think the way they do.
I think this is very important: students listening to people who think differently from them. It’s important to be exposed to different points of view, even when those views are hurtful to you. You want to understand why people think the way they do.
Studyportals: You’ve mentioned the importance of critical thinking, teamwork and practical expertise. As a student looking for a program, how do you recognize those qualities among the hundreds of options online?
Dr. Lino Coria Mendoza: Before I answer, I also want to add creative thinking to that list. Critical thinking, but also creative thinking.
It’s not easy. You look at a bunch of options and need to make a decision. I would say: talk to people from the university. If they’re not willing to talk to you, that would be a red flag. But if there’s someone ready to help and answer your questions, that’s a good start.
When I was a graduate student, I would identify students in the program I wanted to join and email them. Usually, they were happy to answer. It’s like: “I was going to do this work, but now I’ve got an email from some random person. Let me answer it.”
People were really nice to me. I wouldn’t write to hundreds of people. I would identify maybe one person in this program and one in another, and reach out.
Talking to professors can be tricky because I get, I don’t know, 20 emails from potential students every day. Answering in detail is really challenging.
But Northeastern offers seminars and opportunities to talk to me or other professors. I join those sessions, and students are welcome to ask any questions they want about the program.
Co-ops are central to Northeastern’s approach to education, and the university has been a leader in cooperative education for 115 years. With support from academic advisors and co-op coordinators, students gain months of full-time professional experience that connects their classroom learning with their career goals.
That’s why I would say co-op is crucial. This is advice for my own children: go to a program that offers you the opportunity to do at least one co-op.
A co-op and an internship are not quite the same, but if you’re unfamiliar with the term, think of it as an internship where you pause your studies and go to work.
You learn there, but you also bring the knowledge from, let’s say, the first year of your master’s. You do your co-op, return and complete your studies.
I can see our graduates being so successful when they do that. It changes the way they approach everything.
