Student project · Fall 2024

Code Blue AI

AI software that watches for stroke warning signs through a device's camera and microphone and alerts the user or emergency services.

  • Collider Cup XV · Alumni Expo winner
  • Big Ideas Contest grand prize
  • NSF I-Corps participant
  • Berkeley SkyDeck Pad-13 incubator
  • SkyDeck Accelerator ($200K investment)
  • UCSF clinical pilot
  • Represented Berkeley at ACC InVenture Prize
  • UC LAUNCH second prize (shared)
  • Westly Prize finalist
  • U.S. patent issued
  • Seed round open, incl. Republic
  • FieldHealth · AI
  • TeamAshmita Kumar, Ericka Corral
Ashmita Kumar presents on stage beside a slide reading 'Time saved is a brain saved' at Collider Cup XV.
Ashmita Kumar making the case on why time is so important when responding to stroke (Photo by Berkeley SCET / Vicky Liu)

About the project

Code Blue AI is building software that watches for early signs of stroke using the cameras and microphones on phones, computers and TVs that people already own. Doctors can prescribe it to patients at elevated risk. Every 30 seconds it checks speech and facial movement for changes such as slurred words or an uneven expression, deletes the raw images and audio after analysis, and alerts the user or contacts emergency services.

Founder Ashmita Kumar, a Berkeley electrical engineering and computer sciences student, was motivated by her family's experience. Her grandfather's strokes went unrecognized, and a later scare over her father's facial weakness led her to build a first prototype. She took several SCET courses (ENGIN 183A, 183E and 183), and Code Blue AI was named the Alumni Expo winner at SCET's Collider Cup XV in December 2024.

Kumar also won the Big Ideas Contest grand prize, completed the NSF I-Corps program, and joined Berkeley SkyDeck's Pad-13 incubator and then its accelerator track. In April 2025 she represented Berkeley at the ACC InVenture Prize at Notre Dame, and she has since graduated early. The company has been piloting its system with UCSF doctors and plans to seek FDA clearance. As of 2026, its website reports a U.S. patent and results from a 32-patient UCSF pilot, and says it is raising a seed round that includes a community round on Republic.

The team

2 members

Built in

SCET courses
ENGIN 183AFall 2024

A. Richard Newton Lecture Series | A Berkeley Changemaker® Course

ENGIN 183EFall 2024

Technology Entrepreneurship

ENGIN 183Fall 2024

ENGIN 183

See this semester’s courses →