The long, cold “AI winter” is officially over, as both funding and interest in artificial intelligence has increased exponentially around the world.
Now, as its relevance and attraction grows, how are Nearshore IT services vendors finding and preparing talent in order to meet this new wave of AI demand?
Growing from Within
After posing this question to a few providers around the region, one thing became very clear: AI experts are not easily found in the recruitment market, so experienced AI talent generally needs to be developed internally.
“When we started to look for new AI talent, we initially had more than 500 applicants,” said César Hernández, Director of Operations at Intelligens, a Chilean AI development company that was recently picked up by Conversica.
“In the recruitment ad, we didn’t specifically say we were looking for AI specialists, just developers and data scientists who were interested in developing AI solutions. That approach generated a huge level of interest.”
After narrowing down the applicants, the company found that only 10 people already had experience working with AI, such as knowledge of machine learning or Ruby on Rails, for example. However, their main approach for choosing the new team members was to test their problem solving skills, because that, according to Hernández, is hugely important for AI.
“During the first trial month of training, we gave the final applicants one problem solving challenge a week, because, even if people don’t know about AI, those with the right mindset find it easier to learn about it,” he said.
For Argentina’s IT giant Globant, top tier education such as a PhD or a Masters is appreciated, but isn’t a deal breaker when they look for talent – what matters more is a data-driven mindset and a collection of soft skills, such as receptiveness and the ability to bridge the gap between the technical team and the business needs, sharing lessons learned along the way.
“We are seeing a lot of people who have worked on things like data science, statistics, predictive modelling, and machine learning in the past and are now getting back to those roots because there is a market for it,” said Javier Minhondo, VP of Technology – Artificial Intelligence Studio at Globant.
“This kind of talent is ideal, despite not having top-level qualifications. When it comes to experience with certain technologies, it’s the same thing: there’s not just one measure of talent.”
Those who possesses the soft skills and intuition for AI development are valuable, but companies agree that there generally needs to be some level of technical know-how for them to succeed.
“Above all, AI is not one single system or program; it is the implementation of a set of techniques and methodologies that are a subset of programming,” said Luis Derechin, Partner at Nearshore Delivery Solutions. “Therefore, given the right training, smart developers that have the wherewithal for enterprise software will generally be good for AI.”
The profile that NDS looks for is programmers and engineers with a mastery of mathematics, similar to what they would look for in an enterprise developer, except with stronger theoretic knowledge of mathematical logic.
Within the broad scope of what AI entails, it’s not always about just writing or programming algorithms.
“People need to be able to recognize whether a data set is bias or not, or whether the quality of the data is high enough to make some predictions, because data science skills like these are integral to working with AI,” said Minhondo.
“In that sense, we look for skills and capacities related to data science, such as recognizing input and desired output, knowing when to create data proxies, being able to evaluate when a data set is usable or not, or treating data as first class citizens.”
In terms of salaries, market demands have no doubt caused a difference in compensation between specialized data profiles and common, commodity based workers. On average, there are some visible differences, but the salaries are still not close to what is being offered in the US or Silicon Valley.
NDS sees that Mexico doesn’t yet differentiate its AI programmers with higher pay, meaning that the country’s AI professionals currently come at a cheaper rate. “If you were to find someone who had worked on an AI project with an international company, there is no doubt that they would demand a higher salary — and it would have to be paid – but, again, Mexico isn’t quite there yet on the maturity cycle.”
Hernández sees that the Chilean focus on AI is gaining more traction, and the talent pool is exposed to plenty of AI training within local companies, but commented that his Venezuelan team members were initially lacking that same knowledge when they arrived. Even so, due to the situation in the country, Chile has received a lot of its technical talent.
“They were not trained in AI, but were exceptional coders and problem solvers,” he said. “We provided them with the training because they had those essential core skills. If universities were able to create a formal AI program, we would all see a large improvement. Companies like Microsoft are trying to make it a formal part of the curriculum, which is the next step for the region.”
In Mexico, Derechin says, there are currently not enough AI programmers in the country. This is because the cycle of maturity for the market is lagging a few years behind the US, therefore the flux of developers required for these practices hasn’t quite hit the country yet.
“There aren’t enough people that have this specific set of expertise, so we have to create them,” he said. “In a country of 120 million people, there are always a handful of skilled developers who have returned home from other countries, worked with international companies, or those who have studied it in school. Even so, the most important aspect is to create this capability in Mexico – we believe it’s an upcoming field on which companies have to focus.”
The company has between 20 and 30 people working on AI, all 100% Mexican, and is seeing an uptick in AI-related demands in the market, so is finding that more people are gravitating towards the profession.
Developing the Next Wave of AI Talent
Universities in Mexico are aware of the importance of AI, Derechin says, despite academia lagging behind the industry somewhat, so NDS is leveraging its school alliances to push them into teaching the set of technologies that developers should be learning in order to be AI-centric. The company has created an AI course in Tec de Monterrey (ITESM) for that very reason, attracting over 80 students last cycle.
Even for those who have the right profiles, it’s important that they also have coding abilities in something like Python, MATLAB, TensorFlow, C++, or whatever languages are out there. According to Minhondo, data-driven, AI-focused people in Latin America are aware of this need and are adding those skills to their repertoire as they pursue interesting projects to work on.
“At some point, there will be a balance of talent as large companies like IBM, Google, and Microsoft build tools to make life easier to overcome the shortage of data scientists,” said Minhondo. “Educators in the region are also teaching the next wave of AI talent with hands-on projects that can be translated to real-world situations seamlessly, not just peddling theoretical studies.”
In Colombia, specifically, Globant has seen that universities are putting a great deal of focus on data science, which is helping the company to grow its team out there.
“Despite there not being a huge amount of data scientist and AI-related profiles in the region, we are seeing a positive trend in Latin America as more people with these skills sets become available,” said Minhondo.
“It was more difficult five years ago because most of them were only working and researching in an academic capacity, which meant there was a gap in the industry. Today, people who have completed courses in operations, research, and investigation for AI, data science, and machine learning are now working in the market, creating a deeper talent pool, particularly in Argentina and Colombia.”