10-00s
Moritz Wallbrecher
Beyond the Frame: Engineering Human-AI Interaction
Master's Thesis — ongoing
If AI works best when embedded, the next question is whether it needs to be invoked at all. This thesis proposes a system that infers user intent from OS-level signals — clipboard events, dwell patterns, application switching, document metadata — and surfaces contextually appropriate AI affordances before the user articulates a need.
A Computational Intent Pipeline, built on the AI Drop architecture, currently in user research at Kingston University London.
When delay affects something that matters, drivers reach for their phone — not the car. Current in-car systems show route and ETA, but offer no support for the consequence of being late.
Grounded in five contextual interviews, this project explores how an in-car AI can use situational awareness to support drivers during time-critical journeys, without increasing distraction or breaking mental flow. Built as an interactive HMI simulation with scenario-based triggers.
Kingston University London MSc User Experience Design
Focus: Human-AI Interaction, AI-Driven Interfaces, empirical UX research. Expected 2026.
BMW Group Creative Technologist Intern, InCar UX
Built the internal pipeline from Figma and Rive onto the in-car runtime; informed BMW iVentures' investment in Rive.
University of Hertfordshire BA Game Design, 2:1 Honours
Thesis: real-time ray-traced UE5 pixel-streamed BMW i5 configurator, 4K/30fps in-browser.
BMW Group Working Student, HV Electronics
Designed and shipped an internal knowledge platform now used daily by ~140 colleagues across 9 departments.