5-minute Interview Camilo Rodriguez

“ML is a beautiful field because what you can achieve with ML, you cannot achieve with traditional software.”

We’re back with another episode! Today we meet Camilo Rodriguez, founder of the AI consultancy firm MLab.ai.

Tell us a little bit about yourself:

My name is Camilo Rodriguez and I’m based in Paris, France. I have a training and consultancy company in the machine learning field called Machine Learning Lab (MLab.ai). I basically do three things, I teach about AI, ML and Data Science. Then I also consult with companies and I help them deploy ideas and ML projects in their organizations. And the last thing I do is that when I have the time and the energy, I do machine learning for my customers. So I train ML models and I take care of the ML pipelines.

How did you get into the field of ML?:

It was a little bit by accident. I don't actually have an engineering background, I have a sales background. But when I was maybe 30 years old, I converted myself into the software development world for many reasons. I became both a sales and technical guy. So basically I learned how to do software from a web perspective (web development). When you're a freelancer, you have to do everything, so you have to become a full stack developer. So I also got an interest in infrastructure. 

Then I worked in a company for some time that was selling ML solutions from the APIs from Google. It was a consultancy company mostly doing business with the APIs of artificial intelligence as a service. So I discovered the wonderful world of machine learning. I didn't know that you could simply classify an image with just a couple of lines of code by integrating somebody else's APIs. So it was a good discovery for me and I became very curious about that. So I started training myself on machine learning, doing my own machine learning, not being only limited to the API that exists in the market. I understand that there's a lot of people that are really new to ML and AI. There are a lot of misunderstandings, a lot of myths around all this. So part of my job is to train and to explain to people how it works, what can we do with this and what are the resources that you need in a company and so on.

What attracts you to the ML field?:

ML is a beautiful field because what you can achieve with ML, you cannot achieve with traditional software. In one of the exercises that I do when I teach about machine learning, I ask people to explain to me what an apple is. Then I start showing all the differences and all the exceptions to the rule regarding what apples are. ML is a way of doing software that allows you to answer some of the things that are less obvious, like image classification or understanding phrases, and understanding language. In the generative AI field that we have right now in 2024, generating language, generating photos so that it looks real. So I think it's a beautiful field. And I love explaining to people that there are no artificial brains and that it's just software, software that you do with data.

For me, the coolest thing within ML are images, because images are very visual. And even when I teach about machine learning, I start with pictures. This is my first example and the first thing that I show because it's beautiful and it's understandable by anybody. When you start by doing structured data, or when you start showing excel and SQL databases. You cannot explain that. Sure, you can do machine learning with columns and rows of data, but it's a little bit more boring. I mean, the companies need this, of course, but I still love working with images. 

Any resources that you can recommend?:

I have my own YouTube channel where I try to introduce concepts in three or four minute videos. I explain what ML is, like basic concepts for people that are interested in the field, but also more complex subjects as well. Speaking about AI, for me, it's a necessity forever. So speaking about what MLOps is, what is a data pipeline, all this is something very important that people need to understand if you want to be more competitive in today's field. But I can also recommend Andrew Ng. He's one of the fathers of machine learning. So, there are a lot of videos, you can find content from 10-20 years ago. If you're really interested in how machine learning works from a mathematical point of view, I recommend this kind of content for people. 

What's your opinion on MLOps and the broader infrastructure that people need to implement?:

So let me start by first answering, what is MLOps? One thing is to train your machine learning models, another thing is to do wonderful stuff with the data that you have. The problem that most people are not aware of that they're going to have, is that just because you trained a ML model on your computer, the job is not done. This ML model needs to live somewhere in a server, you need to make it available via an API, your model is going to collect data as it's being used, you have to monitor it etc. So MLOps is something that is growing and that requires the attention of the professionals working in this field.