3 min read
We had the opportunity to work together with the nonprofit organization Virufy in the development of a web app which contributes to global public health at the most difficult time in recent history, in the context of the COVID-19 pandemic.
We are proud to be part of this project because we feel that we are contributing our technical expertise to the benefit of the community.
This time we had to study the legislation of each country, such as the General Data Protection Regulation (GDPR), to design the architecture of data storage and management according to the local rules, so that it can be approved and operate without bureaucratic inconveniences.
Virufy is a non-profit research organization that was founded in March 2020, when COVID-19 was declared a pandemic disease.
Initially developed as a Stanford University COVID-19 Response Innovation Lab project and also part of the organization One Young World, Virufy is made up of representatives from all five continents and is motivated by a strong commitment to global health.
Thanks to the work carried out by prestigious institutes such as Carnegie Mellon University, the Massachusetts Institute of Technology and the University of Cambridge, Virufy developed an artificial intelligence algorithm capable of recognizing a COVID-19 infection almost instantly by analyzing cough recordings taken from the microphone of a smartphone or a computer.
The system identifies cough and breathing patterns which are undetectable by the human ear, quickly and reliably, with an effectiveness of over 90%.
Once the pandemic was under control, Virufy focused their efforts on the detection of respiratory diseases in general. Their goal now is to use technology to make prevention free and instantaneous for everyone around the world.
Our programming teams were integrated with Virufy's and we met a group of specialists in very different areas and from all over the world.
Together we developed a Progressive Web App (PWA), which allows to collect cough samples from users from different regions to validate the AI model developed.
Collecting coughs at an early stage was crucial to building a database robust enough to cover a wide range of ethnicities with as many phonological differences as possible. This contributed to the training of the artificial intelligence and reduced the margin of error.
Once the database was built, the system would then compare the samples to determine whether the person who took the test had the disease or not.
By developing the application as a PWA, we were able to remove a barrier. Users do not have to download it from the store, just by logging into the Virufy website, they can donate the sound of their cough, and in the same way perform the tests.
We created a data validation system that also adjusts the AI engine. We designed and deployed the architecture for the information management according to the regulations and laws of the different countries for the use of the information collected.
We used React.JS for the frontend and NodeJS to build the Lambda Functions, since most of the infrastructure is built with serverless microservices. PostgreSQL for the databases, and S3 (AWS) for file storage. We also leverage Github Pages to host the PWA.
If you think we can help your company or you want to join our team, write to us! You can do it to firstname.lastname@example.org or through our social networks.