I spent my first 3 years at Tesla detecting field issues for the vehicle firmware organization. The vehicle firmware data infrastructure was one of the best in the world, collecting signals from over a million vehicles. I switched to the battery data team to better leverage my experience in materials.
Their data infrastructure was much worse than vehicle firmware's and largely a collection of scripts and millions of Excel files. I called my friends from Stanford, now at battery, advanced materials, and semiconductor companies. More than 85 percent told me that their data systems were the same: largely a collection of scripts and spreadsheets. When I asked why they don't use a platform, the answer was that existing platforms are too expensive and hard to use.
That inspired me to start Hypercoordinate, a flexible and affordable system that meets engineers and operators where they are, helping them interpret data, flag and fix materials and chemistry problems instead of managing software infrastructure.