Google DeepMind, a UK-based artificial intelligence (AI) company, is one of the pioneers of AI that is well known today.
The company, led by Demis Hassabis, set the standard for machine learning by building computer programs that taught themselves how to play chess and later Go, considered the most complex board game on the planet.
In its formative years, the DeepMind team was comprised of some of the top minds in the field who were passionate about exploring what many believed was impossible.
These included two South Africans, Shakir Mohamed and Ulrich Paquette.
Mohammed studied electrical and information engineering at the University of the Witwatersrand, where he won the Prime Minister's Medal and the Bernard Price Award.
He also received the Dr. DL Bernstein Award, the Schneider Automation Student Award, and the Altec Electronics Engineering Medal.
After completing his degree with honors in 2005, Mohamed started working at Nedbank in credit analysis and risk analysis for personal mortgages.
But this was just a stepping stone. He received a Commonwealth Scholarship from the University of Cambridge the following year, which enabled him to attend the University to begin a PhD in statistical machine learning.
In a recent interview, he commented on how few people knew what machine learning was at the time, stating that if someone had heard of machine learning, their day would be easier.
His doctoral studies spanned four years and upon completion he moved to Canada to work as a research assistant in the Neural Computation and Adaptive Perception Research Program at the University of British Columbia.
Mr Mohamed then returned to the UK and joined technology start-up DeepMind Technologies in April 2013. The following year, Google acquired the company for $600 million.
Following the acquisition, Mohammed was appointed Research Director and continues to develop general purpose learning algorithms.
His research interests include machine learning, Bayesian statistics, deep learning, socio-technical AI, and broader AI.
According to Google Scholar, his work has currently been cited 37,236 times, nearly 30,000 of which were created after 2020. His h-index is 47.
Ulrich Paquette
Paquet also began his studies in South Africa, where he attended the University of Pretoria, where he says he first became interested in AI.
After completing his master's degree in computer science, he won a Commonwealth scholarship to study at Cambridge. This was the same one that Mohammed received. He also pursued the same field of study as his fellow South Africans.
His doctoral thesis focused on the probabilistic principles of supervised and unsupervised machine learning, which will form the basis of much future research in AI.
After completing his studies, Mr. Paquette took a job at a company that develops technology related to facial recognition and image similarity search.
Shortly thereafter, he joined Microsoft as a senior researcher and was appointed to work on the Xbox One launch team based in the UK and Israel.
Here he developed a system that recommended products based on customer preferences, behavior, and other data.
“We did all the machine learning design for the backend recommender system, and it took about two to three years to get it right,” he said in a 2024 interview.
After finishing his time at Microsoft, he returned to the UK and worked for a company called Vocal IQ, which was acquired by Apple in 2015. As a result, Mr. Paquette was appointed research manager at the American technology giant.
He was there for less than a year before returning to the UK and joining Google DeepMind.
While at DeepMind, Paquette worked on AlphaZero, the company's newest AI project at the time, attempting to create a program that could train itself to master the games of chess, shogi, and Go.
“The way they thought about chess was amazing. It was done in a way that was so different that you could mistake it for creativity,” Magnus Carlsen said on Lex Fridman's podcast.
“The way AlphaZero sacrifices pieces like a knight or two or three pawns is hard to understand and interesting to watch.”
Mohammed and Paquette, who became close friends after studying abroad and later worked together at Google DeepMind, launched Deep Learning Indaba to power AI research in Africa.
The idea arose when the two were attending an AI conference in Barcelona and realized they were the only Africans they met at such gatherings.
Deep Learning Indaba is held annually, with the first edition hosted by the University of the Witwatersrand and the 2025 edition held in Rwanda.
