Headline: Clairvoyant AI (CARA) is a machine learning assistant that provides PSWs and clinicians accurate real time data while predicting responsive behaviors.
Clairvoyant AI (CARA) helps reduce the uncertainty around the challenges professional caregivers and clinicians face with the people they are providing care for and their care network.
Using affordable proprietary wearable devices in a wide variety of styles, CARA learns about an individual’s unique evolving needs and situation. CARA will alert caregivers to adverse events and personalize their monitoring to reliably predict adverse events before they happen and inform their caregiver network.
CARA can objectively know when attention is needed by creating a baseline of bio metric data to learn personalized predictive behavior.
These unique personalized patterns create an evolving specific understanding of the individual to forecast and reduce adverse events in their life and keep their care network up to date and informed on their history and current condition.
Build: - GitHub.com/clairvoyant-app
Programming languages and platform
- Backend: Go, PostgreSQL, Nginx
- iOS: Swift and various open source supporting libraries
- Web frontend: React, chart.js, and various open source supporting libraries (edited)
- Microsoft Band 2 and its API to access sensor data
- Standard Ubuntu Linux LTS
· Varun Grandhi - Co-founder and CTO
· Evan Lee - Backend and iOS developer
· Na Wang - Backend and web developer
· Dominique Dias - Designer
· Terry Myers - Senior Producer
Link to source code:
iOS App Instructions
- Download and install Go on a server
- Pull source code using git
make deps to install dependencies
make to compile
- Run the binary
- Configure nginx as a frontend proxy
- Compile and run the iOS app to start sending data to the backend
Web Frontend Instructions
- Install nodejs
npm install to pull dependencies
npm run build to compile
- Configure nginx to serve compiled output
The full story:
We started with trying to solve problems for two audiences: Professional Service Workers (PSW) and the people they take care of. People being taken care of have such individual challenges that building a standardized solution for every person doesn’t work
Our solution to these challenges was to help people with cognitive challenges through their Professional Service Workers and clincians.
Professional Service Workers build a relationship with each individual and are eager to know their current status and history. What could help Professional Service Workers is knowing in advance when a person under their care may need some extra support.
Using a similar concept as our previous experience in training an emotional response analyzer for a JamesTown Apps client, we built a solution around machine learning from a previously collected data and tweaked it to perform for each individual as opposed to a standard metric.
We currently are under contract with Baycrest hospital developing a Virtual Reality Simulator to explore empathy through VR specifically related to Dementia patients and caregivers. Our team is also waiting on final approval from thet TELUS digital health fund for a companion digital media VR project tied to the upcoming TVO broadcast documentary MUCH TOO YOUNG by film maker Chris Wynn which follows young caregivers as they experience the challenges of their loved ones dealing with the challenges of dementia and Alzheimer’s.
Being aware that complete prevention of accidents for people with dementia is impossible, we also incorporated a system to alert a care network of any responsive behavior or accidents that occur. The potential uses of the data collected are broad and spread across a wide spectrum of caregivers and healthcare professionals.
Researchers, personal caregivers, policy makers, doctors, and specialists will find the data CARA collects useful and insightful. CARA can learn to interpret data collected in a number of ways depending on the user and specific objectives of related parties.