We read the raw sensor stream and hand back signal you can act on: measurements, detections and predictions validated on real patient data and deployed on the line and in the field.
We take on a few problems at a time, where the signal is faint and the domain unforgiving, and we stay until the model holds up in production.
We start from the sensor's raw output, before the device's preprocessing has thrown it away, and we recover it.
Research partnerships are co-developed and co-published with the principal investigator. Technical leadership happens inside the engineering team that owns the sensor.
Founder
Michael spent twenty years building production AI that reads the real world: nine years as a Principal Scientist at Amazon (consumer-health and smart-home devices at Lab126, then the Live Sports AI team at Prime Video) and ten years as Co-Founder and CTO of JustVisual, a visual-search platform that reached 90M monthly active users and $40M ARR. He began in digital signal processing and computer vision for real-time video. Same lesson every decade: a modest sensor with the right model goes further than a great one without.
LinkedInTake a raw physiological, optical or industrial sensor stream from first look to a measurement that holds up in production or a paper that holds up in review. Hands-on signal processing and deep learning alongside clinical and engineering partners. Write to us with a few lines on what you have shipped or published.
Write to usOriginate and structure research partnerships with tier-1 medical centers and technical engagements with companies where the sensor data is the hard part. You should know how hospital research or deep-tech R&D actually moves. Write to us with a few lines on what you have built or closed.
Write to us
MarkerLeap