About Me

I decided to pursue a career in control engineering after excelling in the subject during my undergraduate studies. This academic success led me to join CERN, the European Organization for Nuclear Research, where I was responsible for accelerator control and program development. My time at CERN not only deepened my technical expertise but also reinforced my passion for solving complex engineering problems.

During this period, a pivotal moment reshaped my career trajectory: watching the historic match between AlphaGo and Lee Sedol. This event ignited a profound interest in artificial intelligence, compelling me to explore its vast potential. When I entered graduate school, I found myself captivated by the parallels between human development and the evolution of deep learning models. Both seemed to grow in complexity and capability over time. This realization inspired two personal initiatives: adopting a rigorous exercise routine and delving into research on deep learning model compression.

My research focus shifted to making deep learning more efficient, particularly for tasks like object detection and segmentation in resource-constrained environments. This line of inquiry became more than an academic pursuit—it was a mission to make AI accessible and impactful in real-world applications. I firmly believe that deep learning has the power to drive societal progress, provided it can be effectively deployed in practical settings.

Deep neural networks have already demonstrated superhuman performance in domains like image recognition and strategy games. However, many AI innovations remain confined to research labs, struggling to transition into production environments. As an AI researcher and engineer, I see this as a critical challenge: how to bridge the gap between cutting-edge algorithms and their seamless integration with physical systems.

To address this, I have led projects in computer vision and robotic locomotion, with a strong emphasis on on-device AI. My work has centered on optimizing models for real-time performance and enabling multi-tasking capabilities on resource-limited hardware. These efforts are driven by a clear goal: to develop efficient AI solutions that deliver tangible value, transforming industries and improving lives.

By combining my expertise in control engineering and artificial intelligence, I aim to create systems where AI not only enhances decision-making but also drives action in the physical world. This integrated approach, I believe, is key to unlocking the full potential of AI and ensuring its benefits reach beyond the boundaries of research into the hands of everyday users.