Analog front-end design and a high information-density sampling core for wrist-based electromyography devices, at Facebook Reality Labs — work that shipped in the Meta Neural Band.
Ongoing work designing and testing analog front-end circuits for wrist-worn EMG devices, and a modular sampling platform used to evaluate multiple AFE designs against real research data.
EMG signal quality at the wrist is limited by noise floor, common-mode interference, and the component count needed to condition a very small signal — every design trade-off compounds across a large research fleet.
Prototyped and tested several AFE topologies, driving design decisions that improved signal quality, reduced noise, and cut component count. Built a shared sampling core with high information density so different AFE designs could be evaluated on a common, higher-quality data platform, in collaboration with cross-functional research teams.
A modular evaluation platform now used to accelerate EMG research across multiple concurrent AFE designs, with measurably cleaner signal chains and fewer components per channel. This work shipped in the Meta Neural Band, released alongside Meta Ray-Ban Display glasses.