Brazil loses talent in artificial intelligence

Brazil loses talent in artificial intelligence

Those who have (human) intelligence to develop and train (artificial) intelligences are not unemployed.

For Brazilians, the heated global market opens the way for opportunities abroad, which are more advantageous in financial terms given the devaluation of the real.

Graduated in 2015 at UFSC (Federal University of Santa Catarina), computer engineer Eduardo Bonet moved to Amsterdam, the Netherlands, in 2016.

“The more professional experience, the less the need to have the academic part”, he says. At the time, the professional still did not work directly with the development of artificial intelligence, but managed to migrate from the area in the company where he worked.

Since then, Bonet has changed jobs twice, always looking to work in the field of artificial intelligence.

“I’m a generalist, I don’t have a specific focus. I develop software, work on models, manage projects”, he explains. “Now I work more with software building the infrastructure needed for machine learning [aprendizado de máquina] happen.”

The power of attraction from abroad is also felt in the academic world. “We have difficulty attracting new hires,” says Fernando Osorio, a professor at the ICMC (Institute of Mathematics and Computer Science) at USP in São Carlos and a member of the C4AI (Center for Artificial Intelligence), which is funded by IBM and develops projects related to artificial intelligence in areas of interest in Brazil.

“I have a list of ex-colleagues who were on public service, who were close to retirement, and left for countries like New Zealand, the United States, England”, says Osorio. “They are giving up what in Brazil is considered a security.”

Even doctoral and post-doctoral scholarships, which pay around R$10,000 per month —a rare level in academia—have remained vacant.

According to the Glassdoor platform, which compiles self-declared salary information for thousands of professionals in Brazil, a computer engineer earns an average of R$8,600 per month in the country.


Bonet says that since the beginning of the pandemic, with remote jobs that pay in foreign currency, it has become less advantageous to leave the country.

“Financially, it’s not worth moving,” he says, noting that half of his income goes towards paying taxes.

“It is possible to get opportunities by earning in dollars or euros; in the end, there would be more money left over than moving to another country.”

But the engineer cites other advantages, such as the greater sense of security he has in the Netherlands and better education opportunities for his son, born during the pandemic.

As a disadvantage, Bonet mentions the distance from his family, which would serve as a support network if he and his wife still lived in Brazil. “When we get sick, someone has to take care of our child anyway.”

He didn’t need to learn Dutch to work. “Companies that hire expats usually operate in English,” he says. “Not only that, but companies that require mandatory local language pay less, because the market they are competing for is smaller.”

Osorio, from USP and C4AI, has already considered leaving the country. “Brazil has a characteristic of not investing properly and even destroying the education system, either by an exaggerated low-quality massification or by the destruction of education itself, which seems deliberate, a sabotage”, he criticizes.

“I just don’t go out because I have my parents and my kids,” he continues. “If it were just me, without my family, I would follow these other teachers.”

Mariana Valente, from InternetLab, argues along the same lines: “We have not only an unattractive market, but mainly few incentives to do research in Brazil.”

She highlights that the scenario occurs not only in digital law and impacts the country’s development.

“This undoubtedly has an impact on the topics researched and, ultimately, is a detriment to thinking about technology policies in Brazil”, he continues. “We need to have a local incentive for the production of research, for people to be in the Brazilian market, regardless of the opportunities they offer abroad and how much this can generate some oxygen.”

3 Books to Think About Artificial Intelligence

Shoshana Zuboff, Intrinsic, 800 p.

A professor of sociology at Harvard University, Zuboff is one of the most influential and critical voices of tech giants. In this book, released in 2018, she argues that Google — having massive data collection at the heart of its business model — has ushered in a new era, “in which surveillance is a key mechanism in transforming investments into profits.” “.

Stuart Russell and Peter Norvig, GenLTC, 1016 p.

“We define AI as the study of agents that receive stimuli from the environment and perform actions.” It is with this concept that Russell and Norvig begin the most influential work on the subject, published in 1995 and studied since then in the main universities of the world. In the most recent edition, from 2016, the authors renew analyzes on speech recognition, self-driving cars and the Internet of Things (IoT), areas that have had significant evolution since the 2010s.

  • Superintelligence—Paths, Dangers, Strategies for a New World

Nick Bostrom Darkside Books, 549 p.

Published in 2014, the book explores the evolution of artificial intelligence since the 20th century and projects the scenario that entered the popular imagination with the “Terminator” franchise: what if one day machines are able to develop consciousness and rebel against the humanity? The author is a professor of philosophy at Oxford University and does not put this possibility as the most likely, but develops arguments to keep it on the radar and prevent it from happening.

You May Also Like

Recommended for you

Immediate Peak