I build AI that reads human signals.

Machine learning on human signals — face, eye, gait, voice.

• Mental health sensingDepression, mood and affect read from the phone you already carry.Browse
• Substance use sensingAlcohol and cannabis intoxication from passive sensor data, with just-in-time interventions.Browse
• Clinical movementGait and pose analysis for cerebral palsy and cognitive decline, run on a phone.Browse
• Eye & biometricsIris, pupil and gaze — for identity, and for what the eye gives away.Browse
• HCI & systemsInteraction techniques and the tooling underneath — including the dubbing work behind AstraClips.Browse
• Everything, newest firstFull publication list

The coverage clusters around two stories.

• Depression & eye sensing

  “Depression can be detected simply by looking at a smartphone”

  Also News Medical, Science Daily, Hindustan Times, StudyFinds.org, SciTechDaily, The Daily Guardian, Future Care Capital, 20/20 Magazine, Vision Science Academy.
• Cannabis detection

  “AI can detect if you have recently smoked cannabis”

  Also Phys.org, MobileSyrup, Lab Manager, Greatist, Screenshot Media.

• 06/26Flew to San Francisco to pitch AstraClips at the Deep Tech Demo Day hosted by Third Coast Foundry. Also crossed paths with Fernando De la Torre — my superhero!
• 05/26Gave a guest tutorial on Quantization for Dr. Ahuja's machine learning course at Northwestern.
• 04/26Started as a Q Entrepreneurial Fellow at the Querrey InQbation Lab (Th Q), Northwestern University — building AstraClips.