Researcher's Dream
Motivated by exploration and connection: a career in security R&D and encouraging next-generation talent
Profile
Haruhi Mitani
Data & Security Research Laboratory
Graduate School of Science
Joined Fujitsu in 2023
My Purpose: Contributing to society by leveraging creativity to generate new value beyond conventional boundaries
Article｜2025-10-31
A passion to explore and connect
My dream of becoming a researcher was sparked by an acquaintance, who is a prominent physics researcher. Every time I heard about his latest overseas research, his ability to grasp the essence of complex subjects like particle physics resonated deeply with me, making me think, "How cool is that!" This ignited my aspiration to become a researcher. Influenced by this, I joined the physics research club in high school and engaged in research. The process of formulating my own questions, developing hypotheses and verifying them through experiments was incredibly fascinating, and I became engrossed in exploration.
In addition, due to my parents' work, I had many opportunities to participate in IT-related events and symposiums, which naturally drew my interest towards the field of security. While majoring in physics at university, I also participated in IT-related activities, such as SecHack365, a long-term hackathon designed to cultivate young security innovators, who can tackle security challenges with innovative ideas. This gave me the opportunity to consider the introduction of security education in elementary and junior high schools. Through these experiences, my desire for connecting with people steadily grew, giving me a strong sense of the social significance and technical appeal of the security field. These experiences ultimately motivated me to narrow my job search exclusively to the security sector.
Prioritizing communication and societal connection, not just technology
Among the diverse projects I've worked on since joining Fujitsu, the Secret Leakage Detection Technology project left a particularly strong impression. This involved developing technology that automatically determines whether a document contains confidential information, based on internal information management regulations. I was responsible for the prototype development and accuracy improvement. I struggled with analyzing the causes of result variability and false positives, particularly during the accuracy improvement phase—for instance, when machines judged almost identical content as different, or missed content that should have been detected. It was a continuous process of trial and error. To address these issues, I designed a new evaluation metric with unique weighting to ensure correct detection. As a result, this contributed to reducing the risk of information leakage due to unauthorized secondary use.
Presenting these results at SCIS2025, one of Japan's largest security symposiums, was a valuable experience that made me appreciate my role as a corporate researcher that could deliver real value to society. While preparing the presentation, I wrestled with structuring the story, visualizing the materials and optimizing the delivery. But it was after receiving external evaluations and comments that I made a significant shift from just concentrating on creating technology in R&D to a deeper understanding of the significance of delivering that technology. Through this project, becoming aware of not just the technology itself, but also the power of communication and the connection with society, marked an important turning point for me.
Protecting society's safety and security from unknown threats
In my current research on Multi-AI agent security technology(*1), I consistently develop with the end-user in mind. This technology involves multiple autonomous AI agents collaborating to preempt cyberattacks, as well as supporting countermeasures against vulnerabilities and new threats. Specifically, I am working on developing an AI agent that automatically proposes effective defensive measures against cyberattacks. Attacks are diversifying and becoming more sophisticated every day, with examples including ones designed to make generative AI produce inappropriate responses. Aiming for a system that can respond flexibly and swiftly to even unknown threats, I hope to not only solve technical challenges but also deliver value that contributes to the safety and security of society as a whole. I truly feel the potential for my research to help others, and I feel that I am able to advance my work in a way that is uniquely my own. Looking to the future, I want to help create a world where AI complements human judgment and is utilized practically on the front lines of security.
Taking time to reset and recharge
In my time outside work, I enjoy visiting art museums, going to the movies, and baking bread and sweets. The moments spent quietly contemplating a piece of art or creating something allow me to concentrate in a way that's different from research, and this helps me mentally switch gears. I also love to travel; experiencing scenery and cultures so totally different from my daily life often gives me new perspectives and ideas. I feel that these hobbies provide positive inspiration for my research as well.
Encouraging the next generation of security professionals
I have been interested in supporting people for a long time—so much so that I wrote "I want to be a principal" in my elementary school yearbook—and I also obtained a teaching license. This passion remains with me today, and I want to be involved in developing talent in the security field. The talent shortage has been a long-standing issue in the security industry, and I have begun to think deeply about how to pass the torch to the next generation. For this reason, in addition to my research, I actively participate in a Fujitsu community dedicated to this purpose. At an event called “Girls Meet STEM in Tokyo”, which gives middle and high school girls exposure to STEM careers, I helped inspire participants to think about their future by creating opportunities for them to experience Fujitsu's technology and learn about research-oriented jobs. Through experiences like this, not only was I able to help broaden the career possibilities for these students, but it also served as a reminder of my own original aspirations. Just as past encounters and experiences sparked my decision to enter this field, I now hope to provide moments of inspiration to others, making them think, "Security is interesting," or "I want to know more."
Messages from colleagues
With the evolution of AI technology, the methods of cyberattacks have become increasingly sophisticated and complex. To confront these threats, Haruhi is researching "Multi-AI agent security technology," which is based on the concept of "Active Cyber Defense." The development of new technology is inherently filled with unexpected challenges and difficult decisions, but she overcomes them with her innate creativity and vitality, steadily building a track record of success. She is exceptionally dependable, and we expect her to make even greater strides as a security researcher. (Yuki Unno, Senior Research Director, Data & Security Research Laboratory)
Titles, numerical values, and proper nouns in this document are those reported when this interview was made.
