Eidon Progressive Web App
Multimodal AI data collection at scale
Founding Engineer
As a founding engineer at Eidon AI, built the company's core data collection platform and hardware prototyping pipeline from the ground up.
The platform started as a cross-platform progressive web app for collecting multimodal AI training data at scale. It grew to over 10,000 users who collectively gathered terabytes of image, audio, and video data through task-based 'quests.' Designed systems for real-time ingestion, quality validation, and pipeline processing.
As the company's focus shifted toward embodied AI, I designed and built wearable robotics hardware including an IMU-based arm tracker (shipped 50 beta kits), a BLE HID exoskeleton glove for hand tracking, and POV data capture rigs. Developed quaternion-to-joint-angle math, sensor fusion algorithms, and real-time BLE streaming pipelines.
The work spanned full-stack web, iOS/Android capture apps, embedded firmware, custom PCB design, and 3D-printed hardware, all in service of bridging human motion data to AI training systems.