What research problem are you currently working on, and why is it important?
The most pressing challenge in our field is the reliability and energy efficiency of Edge-AI chips. In recent years, AI-based applications have expanded into various critical safety fields, including public services, communications, transportation, healthcare, and security. This trend places new demands on software and hardware components for greater reliability, computing power, and more efficient use of resources.

Recently, the landscape of AI development has undergone a significant change, characterized by the fusion of AI with edge computing (Edge AI), where data is processed in devices located at the edge of the network. It is predicted that by 2025, over 55% of all data will be processed using Edge AI. This technology brings new opportunities but also requires solutions for limited computational and energy resources, existing AI algorithm limitations, and memory management and access issues.

The reliability of physical chips against operational failures, such as radiation-induced soft errors, nanoelectronics aging, and disturbances from noisy environments, forms the first and most crucial layer of any trustworthy AI technology. Additionally, growing efficiency requirements lead to the need for specialized chips. TalTech is a strong research institution at the forefront of Europe in the reliability and testing of electronic components and systems. Estonia's overall technological potential offers opportunities for the development of specialized chips to support the European Union's ambitions to produce a wider range of powerful chips.

In September, two European projects, TIRAMISU and TAICHIP, will begin under my coordination. Both projects unite top European universities and companies to elevate the reliability and energy efficiency of Edge-AI chips to a new level.

What has been the most socially impactful achievement in your field so far?
Mikk Raud wrote last year in his article "It's Time for Estonia to Realize Its Chip Potential" that chips "form the foundation of today's technology-based lifestyle and the global economy." Since my school years, I have been fascinated by scientists and engineers who, for over half a century, have ensured the development of electronic chips according to Moore's Law. Intel Corporation co-founder Gordon Moore predicted in 1965 that the number of transistors on a chip would double approximately every two years while production costs would remain the same or decrease. This trend holds to this day, resulting in exponential increases in computing power, enabling rapid technological advancements and innovation in all areas of our lives. At the same time, we expect that the growth in computing power will not lead to a decline in the reliability of computational results, and chips based on very complex technologies and very tiny nanotransistors remain the "root of trust."

What concerns you the most about current events in Estonia or the world?
The recent pandemic disrupted multiple supply chains and caused a chip shortage in Europe, which in turn led to the shutdown of production at major car manufacturers. Chris Miller's book "Chip War" excellently describes the current tussle between America and China. Humanity today fears unknown risks associated with rapidly developing artificial intelligence, all occurring against the backdrop of Russia's ongoing military actions against Ukraine. In this situation, the sovereignty of safe and reliable chip technologies in Estonia and Europe is paramount. The European Chip Act, which establishes a strategic fund of 43 billion euros, is a very timely initiative that promotes solutions supporting sovereignty.

How did you get into science, and what draws you to it?
Chance and English language skills. Nearly 20 years ago, I was looking for a topic for my bachelor's thesis. At the same time, university professor Raimund-Johannes Ubar was looking for an interested student for a three-month research trip to Linköping University in Sweden. A key requirement was good English skills, which I had. Later, it turned out that academic Ubar was the creator of a serious school in digital systems testing. The problem of optimizing chip production tests proved very intriguing, and during the three spring months, we conducted research for two conference papers. These formed the basis of my bachelor's thesis. I was amazed that reducing the test time for each chip's physical defects by a few seconds in mass production, while maintaining test reliability, could lower the chip's cost by several percent. I stayed in professor Ubar's lab for the summer, and by the beginning of my master's studies, I already had the results of two more research papers and the core content of my master's thesis. By that time, I knew I would link my life with science, if only for a few years.

What draws me to science is the interaction with very talented, bright-minded people who inspire and constantly teach you something new. The constant rapid development in the field of chips and computing systems also ensures that there's never a dull moment.

What scientific achievement are you particularly proud of?
I am proud to have contributed to the theory of structural decision diagrams created by professor Ubar, which is significant at a global level. This year, his life's work was published by Springer in the book "Structural Decision Diagrams in Digital Test," co-authored by my colleagues professor Jaan Raik, Artur Jutman, and myself. Over the past couple of years, we have managed to form a strong research group to investigate the reliability of AI chips based on neural networks and their models.

What do you hope to achieve in your scientific work?
I agree with the notion that scientific work is similar to elite sports – you have to be the first. Your solution to a scientific problem must be the best among everything that has been offered worldwide, over time. If you are second, the result is not worth publishing. My ambition is to become a top scientist and create a large, sustainable research group that would be a magnet for Estonia's chip-intensive research. Fortunately, there are plenty of research problems in our rapidly developing world.

Conclusion
Maksim Jenihhin's work focuses on enhancing the reliability and energy efficiency of Edge-AI chips, a critical aspect as AI continues to integrate into essential sectors like healthcare and transportation. His achievements, grounded in a robust foundation of Moore's Law and driven by the rapid advancement of AI and edge computing, position him as a pivotal figure in addressing the reliability of physical chips. With significant projects like TIRAMISU and TAICHIP under his coordination, Maksim aims to fortify Europe's position in the global chip industry. His journey into science, sparked by chance and sustained by collaboration with bright minds, continues to inspire his pursuit of excellence in the field. Through his research, Maksim aspires to make Estonia a central hub for innovative chip technology, contributing to a reliable and efficient technological future.

For more information about Maksim Jenihhin, visit his ETIS profile.