What motivates me the most in this work is the intellectual creativity that permeates the scientific process – from the initial idea, through research and development, to publication or integration into products, as well as collaborations between researchers and engineers on large projects
After completing my studies in mechatronics, robotics and automation at the Faculty of Technical Sciences in Novi Sad, as the top student at the University, I wanted to continue my career by exploring advanced robotic systems. This led me to enrol in the EMARO dual master’s programme (European Masters in Advanced Robotics), where I earned my master’s degree in 2013. I had the opportunity to learn from European experts in Italy and France and to work on my master’s project at Japan’s Keio University, where I collaborated with the Japanese Space Agency (JAXA) on mobile robots for planetary exploration.
As an engineer with the RIS group, LAAS-CNRS in Toulouse, I worked on diagnosing rover motion using sequential machine learning models. I later became a research associate with the iBug group at the Department of Computing, Imperial College London, where I applied deep learning methods to recognise human emotions based on multimodal data, including facial expressions and speech.
My career journey continued in the UK, where I completed my Ph.D. at Imperial College London’s Robot Intelligence Lab. There I developed machine learning algorithms for robot control and my thesis focused on neuroevolutionary algorithms that identify a collection of diverse controllers for robots to solve tasks in multiple ways, thereby making it easier for them to adapt to environmental changes.
My goal is to present complex AI topics in an accessible and comprehensible way, inspiring people to engage with artificial intelligence and develop a deeper understanding of its potential and applications
Throughout my career, I’ve transitioned from robotics to computing and artificial intelligence, which required additional effort and self-directed learning. That’s why I believe that a multidisciplinary perspective contributes significantly to the scientific creative process.
I am today a scientist at Google DeepMind, focused on the development and application of large language models. My research interests include reinforcement learning and large language models, such as Gemini. I’m working to advance Gemini’s capabilities, which are implemented across all Google products. What drives me in this field is the intellectual creativity that encompasses the entire scientific process – from conception to research, development and integration into products – along with collaboration on largescale projects.
I’m also dedicated to sharing knowledge of AI through mentoring, scientific education and science popularisation. My goal is to present complex AI topics in an accessible and comprehensible way, inspiring people to engage with artificial intelligence and develop a deeper understanding of its potential and applications.
I believe that the key to success lies in a continuous desire to learn and adapt to change. To anyone aspiring to pursue science, I would say: believe in yourself and your abilities, and remain persistent, because every goal is achievable with effort and dedication.