On Running — Global Retail Interactive is a modular, real-time retail experience developed together with FIELD.IOOpens in a new tab for On Running Opens in a new tab and rolled out across flagship stores, pop-ups, and global events. The system transforms biomechanical running data and in-store movement into generative, brand-defining visualizations on large-scale LED installations—bridging product trial, data analysis, and immersive brand experience.
First launched in New York in 2021, the platform has been deployed worldwide and remains in active use as of 2026.
Roles
- Technical Director
- Lead Unity Developer
- System Architecture & Integration
- Real-Time Graphics & Generative Visuals
- Hardware & Sensor Integration
- Edge AI Integration
- Remote Deployment & Operations Tooling
Stack
- Unity (C#)
- Depth Cameras
- React
- MQTT
- Python
- Edge AI
Outline
Developed in collaboration with FIELD.IOOpens in a new tabOpens in a new tab and multiple technology partners, this project redefined how performance footwear can be experienced in physical retail.
At its core, the installation converts running motion into a living, visual system. Runners are analyzed along a 12-meter track using markerless motion capture, generating a biomechanical profile that is translated into a unique, abstract visualization. Each runner is represented as a weighted mix of six base running types, ensuring that every visual output is distinct and personal rather than categorical.
The system supports multiple deployment modes:
- full runner analysis with real-time skeletal abstraction and replay,
- ambient passerby interaction without active scans,
- autonomous generative “memory” modes that replay prior run signatures when the store is idle.

I was responsible for the technical direction and end-to-end software development of the interactive system. This included the complete Unity application, real-time visualization pipeline, hardware and sensor integration, inter-system communication, and operational tooling.
Beyond implementation, I acted as the technical interface between all stakeholders, defining a scalable infrastructure that allowed the experience to be rolled out globally in different physical and operational configurations—largely via remote deployment.
Over time, the visual language developed for this system became closely associated with On’s retail identity and continues to influence their in-store digital presence worldwide.
System Overview & Technical Characteristics
- Real-time generative visual system
Resolution- and aspect-ratio-independent visuals designed for large-scale, high-density LED walls. The system is fully generative and never repeats. - Motion & presence tracking
Integration of multi-camera setups (top and frontal) for runner skeletal tracking and anonymous passerby detection. Movement subtly influences the visual field even without active scans. - Distributed communication architecture
MQTT-based hub enabling tool-agnostic, low-latency communication between analysis systems, sensors, and the visualization layer. - Edge AI integration
On-device skeletal tracking for real-time runner abstraction, optimized for stability and privacy in retail environments. - Remote operations & control
Web-based control interface (React) allowing store staff to manage brightness, schedules, colorways, branding visibility, screen masking, and campaign-specific configurations from mobile devices. - Global deployment model
Designed for remote installation and maintenance: local teams set up hardware following guidelines, while software deployment, configuration, and updates were handled remotely.
Key Contributions
- Defined and led the technical vision for a scalable, global interactive retail platform
- Architected and implemented the full Unity application, from real-time rendering to hardware integration
- Designed a robust inter-app communication layer using MQTTMessage Queuing Telemetry Transport, an IoT messaging protocol, enabling seamless collaboration between multiple vendors and systems
- Developed runner-specific visual identities, translating biomechanical data into expressive, brand-aligned visuals
- Implemented real-time skeletal abstractions and replay systems for runners
- Built a remote control and operations toolchain for non-technical store staff
- Ensured long-term stability and maintainability across dozens of international deployments
- Trained technical personnel for global rollout and ongoing maintenance
- Contributed to a visual system that shaped and reinforced On Running’s retail identity over multiple years
Why this project matters
On Running — Global Retail Interactive demonstrates how real-time data, computer vision, and generative graphics can transform product trial into a scalable, long-term retail platform. The project bridges biomechanical analysis, spatial interaction, and brand-defining visuals—proving that technically sophisticated systems can operate reliably in global, high-traffic retail environments while remaining flexible, maintainable, and emotionally engaging. It stands as a model for how digital systems can become core infrastructure in physical retail, not temporary spectacle.
















