Álvaro Cabrera is Uruguayan. He studied in the United States and completed a master’s degree at the Mohamed bin Zayed University of Artificial Intelligence in the United Arab Emirates, the world’s first university dedicated entirely to AI. A few months ago, he returned to Uruguay to join one of the startups supported by LAB+: B4-RNA.
Founded by researcher Juan Pablo Tosar, B4-RNA is developing a new diagnostic methodology for the early detection of cancer through a novel class of RNA molecules and a method to detect them in biological fluids such as blood.
How did you end up at the world’s first university for artificial intelligence (AI)?
I’m an industrial engineer from the University of Montevideo. After graduation, I worked for a year at Berkes and another year at Montes del Plata. Then I received a Fulbright scholarship to pursue a master’s in engineering management at Purdue University (United States). Toward the end of the program, the university sent out an email listing international opportunities, and one of them was about the world’s first university dedicated to AI. It had opened about three years earlier and was offering master’s and PhD programs in AI.
The opportunity seemed very interesting to me. First, because I saw artificial intelligence as the future; as an engineer, I felt it was important to learn about it. And second, because I was curious about the country. The UAE has a culture that’s completely different from ours. So I decided to apply. They liked my profile and I was selected. A few months later, I moved to the Emirates.
For your thesis, you worked on applying AI to disease diagnostics. What was the project about?
My supervisor, Eran Segal, is building a database called the Human Phenotype Project. It started with 10,000 participants, now has 30,000, and aims to reach 100,000. The study involves a wide range of medical tests. For example, for two weeks each year, participants must wear a continuous glucose monitor and log everything they eat. On other days, they sleep with devices that track sleep quality. They also undergo DNA, RNA, and genomics testing.
So far, there is extensive data from 30,000 people, each of whom will be followed for 20 years. It’s building a massive database that, when analyzed with AI, could lead to very interesting applications. For example: if someone develops cancer in year 8 of the study, can we see whether earlier years show signs pointing toward the disease? They are also exploring whether, say, changes in voice patterns are linked to disease development.
During my master’s, I worked on that project. Specifically, I wanted to assess whether the database could be combined with information from medical literature (scientific articles, papers, biomedical databases, clinical protocols) to make more robust predictions. We mainly tested this with glucose measurements, blood test data, and retinal images.
We found that when the number of patients studied for a given disease is small, adding literature data improves predictions. But when the dataset is very large, the benefit decreases. So, incorporating medical literature into the analysis would be more useful for rare diseases with very low prevalence.
How did the opportunity to work at B4-RNA come about, and what is your role in the startup?
Some time ago, the rector of the University of Montevideo, Alejandro Cid, reached out to me. He wanted me to join as a PhD student and teach machine learning courses. He also proposed that I become the first doctoral student connected with the university to pursue an industry-applied PhD: working full-time at a company, with that work—plus teaching—counting toward my doctoral degree.
Of course, not just any job fits this type of PhD; it must be research-related. The rector suggested B4-RNA as a good fit, because of the innovative work they’re doing and their need for someone with machine learning expertise.
B4-RNA has a huge amount of data, and they need to analyze it more quickly and automatically with AI. So I reached out to Juan Pablo (Tosar), he liked the idea, and that’s how I joined.
Today, I’m doing my PhD, teaching at the University of Montevideo, and working at B4-RNA. What we’re trying to do is use the RNA sequencing generated by the startup to see whether it’s possible to distinguish healthy individuals from cancer patients, and further subdivide by cancer type and stage.
We don’t develop algorithms from scratch; instead, we fine-tune existing ones—those that have proven to work best for our particular case. This involves a great deal of data processing. The sequencing results from each patient generate around 40 million reads, some of which contain errors or come from sample contamination. Cleaning and processing this data requires an enormous amount of work.
It’s impossible for a person to do it manually without AI, because it means analyzing 40 million sequences per patient.
How do you see the intersection of biotechnology and artificial intelligence today and in the future?
I see it expanding more and more. Data is increasing rapidly, and as datasets grow, AI becomes increasingly useful. If you only need to analyze five blood samples, you might manage by hand. But if you have 10,000 samples or very deep sequencing data, it becomes much harder. The connection between biotech and AI will only keep strengthening. With more data, AI is always more helpful.
In Uruguay, people often talk about “brain drain.” Did returning to contribute to the country influence your decision to come back?
Yes. The main reason I returned was the startup’s focus: cancer and early detection—something I’m very passionate about. I feel that from this area we can generate real impact.
It also mattered that the opportunity was in Uruguay, where I have my family and friends. I lived in the United States and in the Emirates, and the truth is Uruguay is a good place to live. Some things are worse than in other countries, but others are better—especially the people. It’s also very rewarding to contribute to the country by working in a local startup that has the potential to make a global impact.
