Sanibit server
2019
Turns out that healthcare providers need some help with their hand hygiene to reduce hospital-acquired infections. That's where Sanibit came in. Each healthcare provider would wear a wristband, created by our awesome hardware team, that kept track of their hand hygiene whenever they enter or exit a room. Then, those providers and their supervisors could check in on their performance using the Sanibit web app. (That's where I came in.)
For each hand hygiene opportunity, the wristband would record hand hygiene compliance (including not just how long a provider washed their hands, but even whether they donned/doffed the correct PPE!) That compliance data would be periodically uploaded to base stations, which would in turn periodically upload it to the Sanibit server.
The Sanibit server was a headless Django back-end hooked up to a Postgres DB. It was able to efficiently serve the millions of compliance data points to a web app via API in all sorts of formats: average compliance, opportunity count, by hour, by day, by provider, by room, pick your query. It also periodically ran and cached heavier meta-analyses to serve more advanced reports. And finally, it facilitated communication between a front-end and a wristband being set up, via websockets. (Goodness gracious those websockets were the biggest pain in in prod.)
It was deployed -- via GitHub Actions -- to AWS infrastructure which used all of your favorite three-letter acronyms for load balancing, caching, and backups and whatever.
Honestly though? The thing I'm most proud of? Just like with the front-end and the marketing site, I didn't know a thing about this all when I started this job. And by the end? I was building and maintaining all that infrastructure and those wild queries. All pretty darn cool.