Menu
Umi Yamamoto joined Kyulux in November 2023 as an AI/MI Technology Advisor, working closely with the CTO. By June of the following year, he had transitioned to the manager of the Computational R&D group at Kyulux North America, Inc. (KNA), a subsidiary of Kyulux, based in Boston, USA.
What attracted you to Kyulux and made you
decide to join the company?
Is it too high hurdle to aim for such ambitious goals?
I was drawn to Kyulux because I saw an opportunity to delve deeply into Materials Informatics (MI). While leveraging AI in material development—what we commonly refer to as MI—has become somewhat of a standard practice, I believe that MI holds a far greater potential. Specifically, it can drive the creation of entirely new and significant business, not just offer improvements to existing operations. Kyulux, with its foundation in Harvard technology and its ongoing MI efforts since 2016, is currently at a pivotal moment, aiming to create this groundbreaking new business. I felt that taking on this challenge was the best way to truly demonstrate the true potential of MI.
Given that Kyulux aspires to be a global leader in OLED materials development, aiming to be at the top of the MI world is both fitting and necessary. And since I came here to master MI, this is exactly what I signed up for!
How has your impression of the company been
since you joined?
First, I was deeply impressed by the wealth of both experimental and computational data that Kyulux has accumulated. The combination of both types of data is crucial—AI trained solely on experimental data excels at explaining the past, but struggles with future predictions. On the other hand, AI trained only on computational data can often become overly idealistic and detached from reality. Kyulux’s early recognition of this balance and its comprehensive data curation effort provide a unique advantage that’s hard to match elsewhere.
Another surprise was the level of AI technology developed from this data, particularly the high accuracy of properties predictions. I was also struck by the high expectations for precision. From an MI expert’s perspective, this can be challenging, but it’s clear that the feedback and expectations are driven by a genuine belief in the effectiveness of MI and the potential for improvement.
The culture at Kyulux is impressive as well. The solid foundation in MI, or so-called “digital transformation” (DX), is remarkable. Kyumatic™ has become integral to the business systems, with seamless data management and utilization as a given (foundation?). The people at Kyulux are, in a sense, “native” to MI and DX. This gives Kyulux a significant edge, both as a leader in MI and as a materials developer.
What do you mean by “delving into fundamental science”? Can you describe your role at KNA?
When it comes to AI and MI, there’s often an impression of them being black boxeswith a limited understating of their inner workings. However, at Kyulux, we take a different approach. We base our technology development on fundamental science, creating unique features and algorithms grounded in these principles. Being a venture company with a focus on delivering outcomes, Kyulux relies on universal rules—namely, fundamental science—to guide our technology development.
Broadly speaking, my role is to enable Kyulux’s success using the power of MI, and to establish a robust technical foundation for further expansion. This involves developing a specialized DB and AI technology for OLED materials while also considering how these innovations can be applied to other areas. Therefore, I strive to stay abreast of the latest technological trends to lead the development of OLED-specific technologies, while also expanding our technology assets to bridge OLED with other materials. For example, recent advancements in MD (Molecular Dynamics) simulations are a case in point.
Although my role is challenging, it’s also incredibly fulfilling to be at the forefront of the field. It’s beneficial that my broad research experience in various materials fields is being put to good use.
Could you also tell us about some of the standout researchers you’ve worked with?
What aspects of your work do you find
particularly enjoyable?
Additionally, collaborating with highly skilled experimental team members is very rewarding. In MI work, it’s not uncommon for AI to suggest materials that seem difficult to create, but Kyulux’s experimental team excels at synthesizing even unconventional molecules. This hands-on expertise continually enhances the value of our DB.
Furthermore, the thrill of delving into MI and AI is undeniable. On the system side, it’s stimulating to observe US-standard MLOps (Machine-Learning Operations) up close. On the scientific front, engaging in fundamental research alongside American Ph.D. researchers, and diving into core scientific studies is particularly compelling.
Given that Kyulux aspires to be a global leader in OLED materials development, aiming to be at the top of the MI world is both fitting and necessary. And since I came here to master MI, this is exactly what I signed up for!
The most enjoyable part is elevating MI by aligning it with real-world data. While many AI technologies are developed using public databases (DBs), the availability of specific data on features like PLQY (Photo Luminescence Quantum Yield), which customers are really interested in, is quite limited in these public resources. Kyulux, however, conducts its own materials development. This allows us to gather the precise data our customers need, which in turn enables the development of tailored AI solutions. This dual focus on both material and AI development is a unique and exciting aspect of working at Kyulux.
It sounds like a highly specialized and challenging role.
It is, but it’s also incredibly fulfilling to be at the forefront of the field. It’s beneficial that my broad research experience in various materials fields is being put to good use.
What do you hope to achieve at Kyulux?
Can you give a message to those considering
working at Kyulux?
We aim to become a role model for the future of material development, both in Japan and globally. With the demographic shifts in advanced countries, like Japan’s decreasing working-age population, relying on traditional methods and intuition will make it difficult to drive new innovations. Materials Informatics is becoming an essential part of the materials field. Although successful models are still emerging, it’s crucial for cutting-edge materials ventures like Kyulux to lead the way. We are committed to this mission, and we even view it as our social responsibility. I hope that, in the future, our efforts will contribute to a better world, and everyone involved with Kyulux will feel that their efforts were worthwhile.
Kyulux is an exciting place to be! We are using our expertise in both MI and material development to stake our company’s future on MI. This is a unique opportunity to join us at such a pivotal moment. We are fully committed to advancing in this field and achieving success. If you believe in the idea and value of MI, have a deep love for materials and natural science, are eager for serious MI work, or feel that you haven’t fully realized your potential in fundamental science—if any of these resonates with you, please reach out to us. It’s challenging, but together we can shape the future of materials development. Thank you for reading!
