#11 Luis Serrano: Maths PhD, ML at Google, AI educator at Udacity and Apple, AI + Quantum at Zapata Computing
Quantum AI Research Scientist at Zapata Computing (quantum computing co with $67M+ funding), Maths PhD. Ex Lead AI Educator at Apple, ex ML Lead at Udacity and ex ML Engineer at Google
We are Pol Fañanás and Gerard García, two friends passionate and curious about tech, startups and VC sharing weekly high value views from people creating the future. Thanks for reading !
Luis Serrano is a Colombian PhD in Mathematics from the University of Michigan currently working as Quantum Artificial Intelligence (AI) Research Scientist at Zapata Computing, a leading quantum computing company with $67.4M funding named after Emiliano Zapata Salazar (leader of the Mexican revolution), offering quantum solutions to some of the most computationally complex problems including groundbreaking applications in chemistry, finance, logistics, pharmaceuticals, engineering, and materials, among other areas.
Luis is deeply passionate about education and teaching, a true believer that these areas can be improved and that a lot of people should and in fact can be given a better opportunity to learn, fulfill their maximum potential and contribute more to society if we work in better ways to teach and we demystify technical topics that are creating the future so more people can use those concepts to make a bigger positive impact in the world. In this line, Luis teaches machine learning (ML) via learning by doing with simplicity through his book “Grokking Machine Learning”, his ed-tech serrano.academy, and his youtube channel.
Additionally, he has also been involved with multiple social impact/nonprofit projects related to education such as Latinx in AI (community to improve access to AI resources and education for LatinX people) and DreamWakers (nonprofit ed-tech connecting students from low-income backgrounds to career role models).
Previously to going to Canada with Zapata Computing, Luis worked in the San Francisco Bay Area as Lead AI Educator at Apple, ML lead at Udacity, and ML engineer at Google.
Summary
👤 Brief intro: Maths PhD, Google, Udacity, Apple, Zapata
🥇 Win: discovering a passion for education, grateful students
🚫 Fail and lesson: not well in early jobs but eventually finding passion
🚀 Ideal founder: empathic, compassionate, high EQ, doing good
💸 Ideal investor: smart growth, caring about community impact
📈 Markets: ML, AI, quantum, education
🦄 3 startups: Ahura AI, Ubits, Affectiva, Linked AI, Kiwi Bot
👍 3 investors: Zapata, D-wave, Xanadu, Google, IBM, Rigetti, Honeywell
📖 3 books: “Weapons of Math Destruction”, “Algorithms to Live By”, “The Brain That Changes Itself”
Could you give us a brief intro about you and your origins?
I started with mathematics, I enjoyed it and my goal was to be a maths professor. So I did the PhD, post-doc … all the flow. I was close to becoming a professor but I discovered ML and coding and decided to jump into it, after which I went to Silicon Valley.
First, I worked as a programmer at Google and after that I went to Udacity because I was more interested in education and teaching. At Udacity I was the ML lead. Eventually, I switched to Apple doing the same thing as Lead AI Educator where I taught internally about ML.
Ultimately as the ML bug bit me some time ago, I got bit again by another one. The quantum bug. So I joined Zapata, where I am having a lot of fun working in the intersection of AI and Quantum while doing things on the side like writing a book about ML or doing educational videos around machine learning and math through my youtube channel. Education is my passion.
Note from Pol & Gerard: Broadly and inaccurately oversimplifying. ML is an acronym for machine learning, and could be explained as a type of artificial intelligence system which learns from data and makes decisions applying that learning. Quantum is a simplification of quantum computing, which could be defined as an area of computing based on quantum theory that allows for radically better and faster computation.
What would you say has been the biggest win in your life?
I am very blessed that I discovered my passion, the thing I love - teaching. When I found out this one thing and a style that is different and unique, I started doing teaching and sharing material that went on to become somewhat popular, with my channel getting lots of views and subscribers.
I am very happy with the traction and material and especially with getting so many good heartfelt thank you messages from different parts of the world (from the US and Europe to South America, Africa, India and Asia) saying things like “I wanted to become an engineer after learning with your educational content”. That is the biggest win.
Related to the above, and your biggest failure?
This is a great question and it is related to the previous answer. In finding my purpose I had to fail a few times. Now it looks wonderful but at the time it was stressful.
When I was working as a mathematician, I was not doing that well in research. In academia, as a PhD, to stay and thrive you have to be special at it, and I didn't feel that I was doing very special maths research. I was upset it happened, because although my initial goal was to do research, I found that I was enjoying the teaching part much more.
Then I jumped into a programming job at Google but honestly, I do not think I was that good either. A lot of production code, I didn’t enjoy it and I was not doing it very well.
So after those 2 experiences, I dug deep into what is that thing I can do. Thanks to searching for what you really want, being true to yourself, not living someone else dream, and that I was not doing well, I managed to find what I loved and what I should be doing. Thanks to failure because if not I would have stayed frustrated where I was.
What is your ideal founder profile?
I believe there are some stereotypical things that are true such as being fearless, stubborn in a good way, and having good ideas.
But I think other stuff like being empathic, compassionate, and having emotional intelligence is very important too. Because you spend more time with a company than even with your family at times, so the founder there needs to care not just about the product but also about the people. Employees, customers, and everyone. If you don’t care about this is hard to be a founder and I believe you can fall into a negative spiral.
A great founder cares more about impact and people, less about money.
What is your ideal investor profile?
Not as different as the founder.
Yes, there is an aspect of enabling growth and this is important. But it has to be smart growth. Like a brain cell growing while considering the environment and trying to help, not like a cancer which focuses on conquering and not caring about anything else.
You have to grow with the community and care about the overall impact, not just money.
What present and future markets are you most interested in?
Right now especially quantum. Maybe it is obvious because I am involved with this market, but I believe quantum computing and the intersection of AI and quantum is a super interesting area that will continue to grow in terms of hardware, software, ideas, research, and problems (which will get more complicated with time).
Another one really interesting for me, going through a very helpful moment thanks to this 2020, is education. The pandemic made people realize that we have an outdated idea of what education is and we can use tech to really help in critical situations such as not being able to be physically present. Online education in a remote environment is definitely fascinating and although the model still needs to be figured out It is already growing and I hope it grows more.
Could you share with us 3 startups you like and why?
Some interesting ones for me are:
Ahura AI. I am a big fan. They do research on AI and education, trying to really model how humans learn. Not everybody learns the same way and they do a lot of research to understand how to interactively teach people in different ways.
Ubits. A Netflix of education from Colombia, bringing education to companies through a B2B angle.
Affectiva. Really interesting work mixing AI and emotions. They have built an AI that analyzes complex human states in context and provides insights that are used in multiple use cases such as education, gaming, healthcare, and media.
And if it is OK to give a couple more, I would like to add Linked AI (they collect data and use models to generate inputs for other companies to use) and Kiwi Campus (they do robots, with a special focus on delivery, that are really cool).
Could you share with us 3 investors you like and why?
I am not really into investors, but I’d like to mention research companies who are really active on quantum, especially in Canada, a geography that is quite hot in this matter.
Zapata. At Zapata, we are investing in quite a lot of things. Mainly quantum-ready applications for enterprise deployment to address computationally complex problems in use cases like finance, pharmaceuticals, and engineering, among many others.
D-wave. Another interesting player in this quantum realm investing in doing cool stuff regarding the design and manufacturing of quantum computing electronics and services attached.
Xanadu. Also investing significantly in research regarding photonic quantum computing, integrating and designing quantum silicon photonic chips and providing new quantum services too.
Additionally, there are some big companies acting as investors betting hard on quantum like Google, IBM, Rigetti and Honeywell with an effort significantly moving the needle in some aspects. Like when Google confirmed quantum supremacy breakthrough.
What are the 3 books you feel everyone should read and why?
Some that come to my mind are:
“Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy” by Cathy O'Neil. It shows all the problems that AI could do, from a particular model to a particular model and it goes crazy. AI works well but if data is corrupted/biased you can make the problem bigger. You have to understand all the things you have to be careful with. You need to be aware of this ethical angle.
“Algorithms to Live By: The Computer Science of Human Decisions” by Brian Christian. It explains is very good how the insights of computer algorithms can be applied helping to solve common decision-making problems and helping to understand how the mind works. I really enjoyed it.
“The Brain That Changes Itself: Stories of Personal Triumph from the Frontiers of Brain Science” by Norman Doidge. One that is not so much about tech but about neuroscience with a bunch of specific examples about neuroplasticity, explaining the power of thought from a scientific way and illuminating how people with mental limitations or brain damage were successfully treated or cured by thinking positively in a certain way, eventually changing their brain structure. Amazing to learn about how you can rewire your brain and help yourself.
I would like to also add “How Not To Be Wrong. The Power of Mathematical Thinking” by Jordan Ellenberg. A bunch of interesting fallacies and how we can think in a different way and use maths to see the real structure under the messy surface of the world.
Note from Pol & Gerard: We would also like to highlight the book “Grokking Machine Learning” by Luis Serrano himself, which in a great fashion aims to dispel the myth that machine learning (ML) is difficult by focusing on how you can do cool projects (from Spam detection to language analysis and image recognition), using only standard Python code and high school level math, leveraging ML tools already available and learning by doing.
WILDCARD QUESTION
About AI & quantum computing intersection, could you help us pull the curtain by giving us a brief deep dive in what is the current stage of innovation and what could we see in the future? About education, how do you think it has to evolve so that a significant number of the population is not left behind and is empowered to create value in a world increasingly dominated by technology?
Great question !
About the 1st part. Quantum is going to make a huge impact in many technical fields because it can improve many algorithms that are very slow like factoring numbers and then it can break another previous stuff like crypto. And specifically, ML is one of the places where quantum can make huge impact in both directions, you can solve ML with quantum and you can solve quantum with ML.
One of the first things where impact can be huge is in generative learning. Quantum can help find distributions and generate data better and faster. Picking it out of intractable distributions is hard, but it may be very easy for a quantum system. For instance the opposite (regular supervised learning) can take the image of a dog and it tells you it is a dog. For a generative model, instead, you tell it to draw a dog, and it draws a dog. This is one area where we believe we could do a lot of improvement because many distributions are hard to sample. Taking random samples is not easy whereas in quantum it is easy because to get that random angle with quantum computing you just need to observe. A lot of things are much more natural to a quantum algorithm and we would like to explore all of that. Not just the speed but the performance. And this was just one quick easy example but there is impact in many thinks like finance, chemistry, engineering. Almost anything.
About the 2nd part, education has to evolve? Absolutely. Education needs to change. We need to demystify technical stuff, remove the formulas and introduce concepts. Make sure people are not afraid of mathematics and programming. Right now really impactful stuff like AI and quantum is selective. Only people who can handle abstract thinking via formulas and code can thrive. But we are shooting ourselves in the foot, as there are many different minded people that could contribute but don’t necessarily handle abstraction very well.
I believe there are 4 particular ways (at least) that Education needs to for proper evolution:
Geographic. Every person, everywhere.
Socioeconomic. If you can’t afford it you will still get it and contribute to society.
Chronologic. If you are 10 or 90, it does not matter, nobody should stop you from getting education. It is a bad idea to learn at the beginning of your life and be expected to produce the rest of it. It should not be like that.
Pedagogic (type of learning). Everybody learns in a different way. Tech? Who can think in abstraction and handle formulas. But what if not everybody can? We have smart people thinking in a different way and we need those people. We need to teach for everybody but not in a standard rigid style. Also more interactive, not passive. For instance, you tend to become smarter when you are breaking something apart and learning by doing, not sitting there passively and feeling stupid.
The goal is for every human to reach her/his particular potential, but we only have a fraction of them doing it. It reminds me of philosophy, in the sense of thinking and writing it down. One thing is the concept (thinking), the other is the language (writing). Why forcing people to write in a specific language in order to be able to think properly and if they don’t know the language they can’t think? You can be a non abstract person and also have the potential to be really helpful in AI or anything technical, even though you can’t manage a specific type of high level language. Everything should be approachable by everyone. And specifically stuff like AI needs more people to understand the problem and solve it, in order to evolve. Imagine a world where everyone could learn and reach their full potential.
Big thanks Luis for sharing your views with us !
Big thanks to you, reader, for your time and interest !
If you enjoyed it, subscribe and we’ll be back with you next week. 🙏🏼