Mercedes Research
In-cabin AI / UX research and prototyping for Mercedes-Benz — four shipped research prototypes from MBRDNA Sunnyvale AI Labs and the Carlsbad Design Studio, 2023–2026.
Conner Ward led in-cabin AI and UX research for Mercedes-Benz at MBRDNA's Sunnyvale AI Labs and the Carlsbad Design Studio, building a set of research prototypes spanning LLM-disambiguated voice intent, world-anchored driver AR, and generative in-cabin media.
This is the group index for that body of work (2023–2026): the FuzzyInput LLM-disambiguated intent UI, the DriverAR world-anchored navigation experiment, and the Album Outpainting generative-display pipeline. Each project is its own page, covering the problem, the approach, Conner's role, and the outcome at a portfolio level of detail.
-
01
FuzzyInput
2024–2025Sunnyvale AI LabsLLM-disambiguated in-cabin intent UI. Telemetry across 21M+ daily transactions, wizard-of-Oz cabin-buck studies, Android + Arduino prototype, executive HU car demo.
-
02
DriverAR
2024Carlsbad Design StudioRealtime 3DOF in-car head tracking for an XReal AR driving experience. Sensor fusion across Zed stereo, IMU, and MediaPipe into a Unity render.
-
03
Album Outpainting at scale
2024–2025Sunnyvale AI LabsGenerative outpainting of album cover art to fill Mercedes' pillar-to-pillar displays. A cached, fleet-scale oneAI pipeline — LLM-prompted, ComfyUI-generated — gated for driver distraction.
FAQ
What is Mercedes Research?
Mercedes Research is the portfolio group collecting Conner Ward's in-cabin AI and UX research and prototyping for Mercedes-Benz, done at MBRDNA's Sunnyvale AI Labs and the Carlsbad Design Studio (2023–2026). It spans three projects: FuzzyInput (LLM-disambiguated intent UI), DriverAR (world-anchored in-car navigation AR), and Album Outpainting (generative album-art backgrounds for the cabin displays).
What was Conner Ward's role?
Across the Mercedes-Benz research projects, Conner Ward worked as a UX engineer / researcher and individual contributor with cross-team scope — covering telemetry and user research, prototyping, and the engineering pipelines, working with design, ML, head-unit platform, and hardware teams in the US and Germany.
What technologies and methods were used?
The Mercedes-Benz research work combined fleet telemetry analysis, wizard-of-Oz cabin-buck studies with gaze tracking, LLM-based intent disambiguation, sensor-fusion head tracking (stereo camera, IMU, face landmarks) rendered in Unity, and a generative outpainting pipeline (ComfyUI diffusion plus LLM prompting) on Mercedes' oneAI platform, prototyped on Android and production head-unit hardware.