Takafumi Horiuchi | Design engineer
Recent self-initiated work
-
-
-
-
-
Personalizing interfaces through constraint-driven design (August 2026)
I explored how to generate interfaces tailored to cognitive, sensory, and physical characteristics within the limits of scientific evidence and design constraints. I used a car-navigation proof of concept to generate interfaces and check them against the requirements, exploring the approach’s potential and limitations. Usability and safety have not been evaluated with users.
-
Generating a one-of-a-kind Pokémon (July 2026)
A research prototype that generates a new Pokémon species from the player’s words during play, then lets them encounter and catch it. I designed the constraints the AI must respect throughout that experience.
Earlier academic work
-
Master’s thesis on human–robot interaction design (July 2021)
I proposed and implemented a collision-prevention mechanism that uses the operator’s field of view estimated from head orientation to help determine where a robot may move, without requiring objects to be registered in advance. I confirmed the intended behavior in five demonstration trials.
-
A VR system for children to prototype interactive objects (February 2021)
A VR prototyping environment in which children can assemble and run the smart objects they imagine in the SNaP board game. They define each object’s conditions and behavior using three cards: environment, input, and output.
-
Motion analysis of a piano action (January 2020)
This project reconstructed a piano action in 3D from a cross-sectional drawing and used theoretical calculations and SolidWorks analysis to examine how input at the key translates into hammer velocity and the string’s reaction force.
-
Three design studio projects (2019–2020)
Three group product-design projects completed over three semesters at Politecnico di Milano: a coffee machine for shared spaces, a powered air-purifying respirator (PAPR) for welders, and a contactless blood-donation screening device.
-
Bachelor’s thesis on constraint programming (June 2019)
I developed a method that identifies conditions that can no longer become true during a simulation and removes them from the search. It reduced execution time by roughly half in the evaluation model without changing the simulation result.