Hamza Merzić, born in 1993, is one of the artificial intelligence researchers at Google DeepMind in London, as well as one of the founders and a member of the Board of Directors of the Association for the Advancement of Science and Technology (ANNT). He tells Klix.ba how he made it to Google, one of the world's largest internet companies.

This Sarajevo native graduated from the Department of Automatic Control and Electronics at the Faculty of Electrical Engineering, University of Sarajevo, in 2015 as the best student of his year and a recipient of the Golden Badge. During his studies he served as vice-president of the IEEE Student Branch in Sarajevo, in which capacity he worked on projects such as the development of an autonomous photovoltaic system (a solar tree) and organised robotics workshops for secondary school and university students.

He completed his master's degree in robotics and artificial intelligence in 2018 at ETH Zurich in Switzerland, in collaboration with the Max Planck Institute for Intelligent Systems in Tübingen, Germany. During his master's studies he also gained experience in the start-up world, working for Rapyuta Robotics, a Japanese autonomous robotics company. He begins by telling us that mathematics was always one of his greatest passions and that for a long time he believed it would be his career.

"Although I wanted to focus exclusively on mathematics, towards the end of secondary school I realised that I wanted to do something applied after all, but I by no means wanted to give up on this science. I decided to enrol at the Faculty of Electrical Engineering in Sarajevo, in the Automatic Control and Electronics programme. That programme is known for demanding a lot of mathematics, and what attracted me even more was the robotics course. During my studies in Sarajevo I spent a lot of time taking online courses in various fields, and that is how I came across artificial intelligence. I remember how fascinated I was the first time I read about learning algorithms, which, like children, learn on the principle of rewards and punishments, unlike classical programs where every behaviour is explicitly defined," said Merzić.

A focus on robotics and artificial intelligence

He says that at the time he was studying, modern artificial intelligence was at an early stage, so he soon shifted his focus back to robotics. He completed his undergraduate studies and enrolled in a master's programme in robotics (Robotics, Systems and Control) at ETH Zurich. In the second year of his master's he took a course in artificial intelligence and, he tells us, was simply astonished at how far the field had advanced; the same fascination he had felt a few years earlier was reawakened, and he decided to learn more about the field.

"For my master's thesis I took on a project at the intersection of robotics and artificial intelligence and began planning my next steps. My dream became doing scientific research at one of the leading companies such as Google DeepMind or OpenAI. However, the requirement for most such positions was a PhD, which did not particularly appeal to me, as I felt that a PhD was not the most efficient way of doing science, especially in fields like artificial intelligence, which require a great deal of resources and infrastructure that few universities and institutes can afford. Nevertheless, I came to terms with the fact that I would need to spend at least another three years in academia to build the career I wanted. In the same period, one of the professors in Zurich, who had offered me a PhD under his supervision, told me that he was leaving his professorship to start working as a researcher at DeepMind. Luckily for me, he offered me a letter of recommendation for a research position, which of course I accepted, and so, together with several other PhD students from his lab, I began the interviews," he added.

He points out that at first he was not very optimistic: he had only started seriously learning about artificial intelligence a semester earlier, and he had already applied twice for a summer student internship at Google and been rejected both times. However, the interviews went well one after another, and he soon realised that the only ones left from the lab were a final-year PhD student and himself.

"That was one of the most intense periods of my life, because I was simultaneously finishing my master's thesis and interviewing for several companies and PhD positions. The whole interview process for DeepMind took three and a half months in total, and during that time I had interviews with 14 people. I received the job offer on 18 December the year before last and I must admit I was overjoyed, especially because I felt that I had somehow skipped the PhD and yet still got the chance to do science," Hamza Merzić explains.

Working on "testing ideas"

"A working day at DeepMind is very particular and quite different from other companies, even from other parts of Google. Scientists like to sleep, so our nominal working hours are from 10 am to 6 pm, but even that is very flexible and you can often catch me arriving at work in the afternoon. Working outside the office is not a problem either, as long as the work gets done on time. In addition, we have all of Google's benefits, such as free food in the office, gym access, massages and so on. As for the work itself, the dynamic is more like a research institute than a conventional company. We are first and foremost engaged in scientific research, so our time is divided between independent learning, reading scientific papers, implementing algorithms, running and analysing experiments, and presenting results. We present most of our results externally as well, through scientific publications and blog posts," the young researcher tells us.

What he likes most, he explains, is that he and his colleagues do not develop products, so their deadlines are not tied to delivering finished products but to "testing ideas". If an idea does not work – which, he tells us with a laugh, is mostly the case – that is still a success, because every such attempt is followed by a valuable lesson.

"Google DeepMind is divided into two parts: a research part and an applied part. The idea behind this is that people doing science are not incentivised to work on profitable ideas, but have complete freedom to develop their own ideas, which then reach the applied part of the company that deals with potential applications. Of course, when I say complete freedom, I must stress that all our work in some way aligns with our mission of developing artificial general intelligence and, as such, must be approved by an ethics board in order to avoid malicious applications," Merzić points out.

Young people from BiH are hungry for knowledge and challenges

When we asked him what problems young people in BiH currently face and whether they really have opportunities to advance, he explained that there are many obstacles young people from BiH need to overcome, from unemployment and poverty to insufficient investment in education, and so on. He also stresses that it is hard for young people to find the courage to study abroad, which for most is impossible without scholarships.

"Despite all the hardships, we still have many talented young people hungry for knowledge and challenges. Indeed, there are so many examples of young people winning medals at the world's most prestigious science competitions or achieving significant results in their studies abroad. I think young people have the opportunity to progress up to a certain level and that our education system up to postgraduate studies is still of good quality and can offer a lot. Still, I must admit that when I applied to study abroad, I too was afraid that people did not have a good opinion of our universities, but I was proven wrong when I received positive responses from most of the universities I had applied to. Unfortunately, it is perhaps even harder to come back, especially for our highly educated diaspora, because, sadly, there is no industry in which such people could find a place. We all know about these problems, but the real question is what we can do to improve the situation in BiH," says the 26-year-old from Sarajevo.

Finally, we asked him whether he could draw a parallel between life in Sarajevo and London:

"Life in London is much more hectic; everything is somehow far away because the city is huge, so it is not unusual for people to live and work in London and still need an hour to get to work. I think bigger cities make people more distant; there are far fewer spontaneous get-togethers, so people are lonelier as a result. But on the other hand, London is very exciting; there is always something to do, lots of parties, exhibitions, fantastic museums and galleries, and art in general has an important place in the everyday life of 'Londoners'. People have more room to express themselves, there is a different kind of respect, and everyone has the chance to be who they are and live as they feel. Both cities have access to beautiful nature; London is very rich in parks and recognised long ago the importance of green spaces in a city. The food is also really excellent and very diverse, although unfortunately I have not yet managed to find good ćevapi. Perhaps my favourite thing is London's excellent connections and cheap flights to lots of destinations, so you will rarely find me here at weekends," concludes Hamza Merzić, a Bosnian-Herzegovinian researcher working at Google DeepMind. Source: klix.ba