Dongfeng/Huawei

Ground the product vision.
Build the capability to deliver it.

Future cockpit concepts needed a product direction before there was a fixed roadmap or feature list. Interior, exterior, color and materials, and UX/UI had to work from a shared understanding of customer needs, while teams in Wuhan and Shanghai considered expectations beyond the Chinese market.

Creative production also became a bottleneck. Concept videos moved through sequential handoffs across disciplines, taking 2–4 weeks. The challenge was to connect research with product decisions and give teams a faster, repeatable way to explore and communicate those decisions.

Frame the research.
Implement the workflow.

I set the experience vision and design principles, structured a research partnership with Hunan University and framed the questions it needed to answer. I translated between researchers and product teams, then synthesized findings into opportunity areas and principles the design disciplines could use together.

Across Wuhan and Shanghai, I provided creative, strategic and educational leadership to 43 colleagues. Written principles, shared design reviews and early involvement from product and engineering helped make the direction usable across the teams. Alongside project reviews, I developed and taught a structured mentoring program covering UX/UI, research, motion and Agile delivery, connecting design principles with practical project decisions.

I built an MCP-orchestrated AI production workflow and taught it through live demonstrations. Teams adopted and implemented the method, with clear confidentiality rules and expert review of AI output, craft and correctness.

Research into product direction.
AI into team productivity.

The research informed future cockpit experience principles and four flagship concepts validated through VR and sound prototypes. The concept-video production workflow achieved 10–20× faster production, moving from 2–4 weeks of multidisciplinary handoffs to one designer running the workflow in one day.

My wider contribution included customer journeys, shared character references, cockpit and projection studies, and creative direction for Dongfeng/Huawei vehicle programs. The character development and motion pipeline pages show those applications; the collection includes concepts and development studies.

Hunan University.
Research into product direction.

My role. I structured the partnership with Hunan University’s PhD-led research center, framed the questions the study needed to answer and translated between academia and the product teams. I synthesized findings into opportunity areas and design principles for future cockpit experiences.

Research scale and methods. The study combined a big-data layer of 1.78M+ user records with 30 in-depth interviews with family users across 11 cities. In-vehicle fieldwork included ride-alongs, day reconstruction and co-creation workshops, with LLM-assisted analysis during synthesis.

Product direction and outcome. Findings informed two family personas and a Kano-graded scenario library, with pain points prioritized by frequency and emotional value. The process narrowed 375 ideas into 100 innovation cards and four flagship concepts, validated through VR and sound prototypes. These findings gave interior, exterior, color and materials, and UX/UI a shared basis for experience decisions.

An adopted AI workflow.
10–20× faster creative production.

My role. I identified concept-video production as a bottleneck, built the workflow, demonstrated its value and taught it to Dongfeng’s teams. They adopted and implemented the method in their work.

The workflow. I used MCP to connect AI models directly with After Effects, Premiere, Blender and Unreal, alongside generative image, video and sound tools. One designer could carry the brief through concepts, storytelling, frames, prompts, images, video and sound, reducing the sequential handoffs between disciplines.

Measured impact. Concept-video production moved from a multidisciplinary team taking 2–4 weeks to one designer running the workflow in one day, achieving 10–20× faster production. The team could explore more directions within a sprint. The comparison applies to this workflow; timing varies with the brief, reviews and finishing requirements.

Quality and adoption. Live demonstrations and teaching helped colleagues learn the method and retain ownership of the design. I established confidentiality rules, treated AI-generated claims as material to verify and kept expert review of craft and correctness. Shared design references, motion presets and sound systems kept the outputs connected to the intended design language.