Symbionic: Bionic Prosthetic Arm

Overview

Co-founded Symbionic, a Chennai-based startup building affordable bionic prosthetic arms for users in India and across South Asia. The core problem we tackled: how do you translate noisy, real-time EEG biopotentials into reliable hand gesture commands that a prosthetic arm can act on within the latency budget of a natural motion?

The work

I owned the embedded signal-processing layer of the bionic arm. EEG sensors picked up raw biopotentials from the user. My code on the embedded controller filtered and post-processed the signal, then ran a hand gesture recognition (HGR) model that mapped clean signal patterns to specific gestures (grip, point, open, etc.). The arm responded to recognized gestures in real time.

The hard part was the signal processing pipeline: EEG is noisy, the latency budget is tight (a delayed prosthetic doesn’t feel like your arm), and the embedded hardware can’t run the same models a laptop can. We iterated on filter chains and lightweight classifiers until the responsiveness felt natural to early users.

Outcomes

The work was funded by India’s Department of Science & Technology. The company continues to operate; the most current information lives on the Symbionic site.