Bridging AI, HCI & interactive systems.
Multidisciplinary engineer with a foundation in artificial intelligence, human-machine interaction (HMI/HCI), and advanced display systems. Based in Tokyo, JP with global industry and research exposure across Japan, Germany, Netherlands, France and Mongolia.
Chronology & Evolution
Tracing the path from low-resource linguistic neural modeling to physical autonomous machines.
Research Engineer at BMW Tech Office JP — AI, Advanced UI/UX & Interactive Systems
Technical lead for UI/UX Concept Development of APAC Innovation Car (AIC) DualHUD and MINI APAC Innovation Car (MAIC), orchestrating international internal and external teams for integrated technology user journeys. Designed dynamic projection UI/UX concepts with distortion masking, multi-purpose gesture control for infotainment, AR Head-Up Display concepts for challenging windshields, and refined motion graphics for Munich showcases. Conducting research on domain-specific LLM evaluation and embedded edge-based small language models.
Master of Engineering: Engineering Science & Design (Minor: AI & Data Science)
Tonen International Foundation Scholar. Conducting research on Parameter-Efficient LLM fine-tuning and multimodal foundation models. Master thesis on multimodal affect recognition (Affect-LLaMA accepted to BMW Summer School 2025; poster at Eurocom AI Conference 2024).
Founding Member & Data Scientist: SensUs 2024 Competition
Developed a real-time, continuous monitoring biosensor for kidney disease utilizing creatinine as a biomarker. Contributed to data science models and business development, achieving 1st Place in the Business Translation Award and 2nd Place in the Innovation Award at SensUs 2024.
Cross Domain Computing Solutions — ADAS & Computer Vision
Contributed to technical project management for the collaboration between Bosch & Toyota Woven City ADAS tech stack development. Tested and validated computer vision modules (object detection, lane detection, etc.), dataset validation, and model quality assurance.
Bachelor of Engineering: Transdisciplinary Science & Engineering (Valedictorian)
MEXT Japanese Government Scholar, Valedictorian of the 2024 Graduating Class, and recipient of the Best Independent Research Award. Investigated low-resource morphological image captioning for Mongolian using ResNet convolutional features and recurrent attention decoders with Byte-Pair Encoding.
Escape from the Pure Simulation
Artificial intelligence confined to cloud servers operates in a sterile, frictionless environment. It assumes instantaneous communication, infinite compute, and abstract loss landscapes. Yet the real world does not operate on tensors alone.
When machine learning encounters physical hardware, it confronts real friction: sensor noise, motor gear backlash, latency, microsecond timer interrupts, gravitational deflection, and heat dissipation. Building custom parametric carbon-fiber brackets, tuning closed-loop PID control loops on microcontrollers, and capturing the movement with macro lenses grounds technical inquiry in physical truth.