Miroslav K.

An app for treatment of sleep disorders and improving sleep quality. I processed signals from iPhone sensors: microphone (sound pressure), accelerometer (rapid movement detection), and apple watch sensors: heart rate and accelerometer. I also did an acoustic calibration of the microphone and output device in the app, because the user could use both internal and external speakers, bluetooth speakers, headphones, bluetooth headphones, airpods, airplay, etc. The microphone could also be damaged or clogged with a brothel, which affects acoustics and the resulting therapeutic properties of the product. We streamed the data from apple watch using watch connectivity and sent it from iOS via websocket to the server. On the server, machine learning was used to evaluate which sound and how loud to play, and this information was streamed back to the app. So the app was kind of a DJ, the transitions between sounds had to be solved, the movement of the environment was masked so that the patient would not wake up and so on. The data had to be cached well and dealt with a lot of states and combinations of states, because the connection to the apple watch, headphones, internet, etc. did not always work. The user could get a phone or an alarm, etc. And all in the highest security standard.

I was leading a team of three in an international environment, working with security technologies, writing Swift code, messaging with Socket.io. We’re developing a sophisticated social network. My key responsibility was to ensure quality codebase and fast delivery in a tight time schedule. Deployments are done through Bitrise CI, task management in Trello. Fast iterations and regular calls with client ensure the project always runs in the right direction. As our client was a health startup, one of my responsibilities was to ensure security and privacy when handling sensitive information.