Sergey M.

Profusa is a company, creating tissue-integrating biosensors for continuous contactless monitoring of blood chemistries. Some of the beneficiaries of Profusa include athletes, soldiers, and diabetics. Medical institutions are also using Profusa sensors for live monitoring the oxygen and sugar levels. In order to fine-tune sensors during development, Profusa needed a system for real-time data collection, analysis, and consolidation. Their legacy system was not adequate to handle the task, so my team and I took charge of their legacy system modernization. I was a Manager and a Team Lead for this project. Having started with only 3 people on the team, I was able to made thorough analysis of all issues that affected system performance and stability. The challenge was to collect data from multiple sources like Bluetooth sensors, USB Sensors, and CSV-files. Since every source had it’s own data structure, the first thing we had to implement was a data consolidation algorithm. As a result, no matter the input source, all of the data had been converted to a single internal format. This significantly reduced code duplication and improved maintainability. Performance and stability of chart rendering was another issue we had addressed. On large data sets, the chart was freezing, while on medium data sets, it would take a number of minutes for the chart to display. Benchmark clearly showed the problem existed on multiple levels. We had replaced the charting component, enabled the network traffic compression, and completely reworked the data pre-processing algorithms. This increased chart display speed by 40 times and resolved all known stability issues.