Our inspiration stems from an on-going struggle with providing adequate PPEs for myself (a front-line healthcare worker), my colleagues, and the public. There is an blaring inefficiency present that needs to be addressed. As allergy season approaches, we run the risk of misinterpreting a runny nose from pollen for CoVID-19 causing us to needlessly expend scarce resources such as PPEs and test kits. Our healthcare system is struggling even prior to this pandemic, it cannot take another blow like this without innovation.

What it does

Our solution relies on AI analysis diseases that have similar clinical presentation to the naked eye. This allows us to deferentially diagnose CoVID-19, flu, cold, allergies, and etc., thus supporting front-line healthcare workers in deciding when to dawn costly PPEs and when to use scarce test kits.

How I built it

We built it by quickly re-purposing a previous platform that tackled the social determinants of health in a similar way. By leveraging the classification capabilities of Tensorflow and a simple submission form for input, we were able to make simple predictions about the disease simply by taking a patient's symptoms.

Challenges I ran into

It felt like everything I did was a challenge. That was because I'm not formally trained in anything related to coding and design. That being said, each challenge served as motivation to learn more. It was a very humbling experience.

Accomplishments that I'm proud of

I'm proud of all the design principles I've picked up along the way. From never thinking twice about the size of button, to trying to get the exact pixel size for different elements. It will never cease to amaze me how much thought is put into such a seemingly minute detail.

What I learned

I learnt that rome was not built in a day. Every time I felt like something is done, I find a new glitch or thing to improve on. Again, very humbling.

What's next for Virus Engine

We are collecting more data points to feed the AI. No other way to do it. We need to collect thousands of examples of what CoVID-19 looks like for each patient and train the machine.

Try It out


corvid, javascript, tensorflow

Devpost Software Identifier