04 Feb Video: Presentation of the GUI of Aigine
The good thing.
The work needed to perform a personal data inventory is reduced with more than 97% by combining artificial intelligence, modern workflow and an intuitive user experience. At the same time, data subject requests become much easier.
The magic in the background.
Simplifying the user interface also means that the actual magic happens in the background. So here are some of the cool tech things that are never visible:
We are using StoredIQ from IBM and AI Miner from Elinar to offer a hybrid cloud/on-site SaaS, that generates anonymized annotation code locally with Watson NLU. Actual AI training is done centrally using IBM Power HPC appliances. Each Appliance has 20-36 Power8/9 cores and 4 high end NVidia GPUs connected directly to Power CPUs using Nvidia proprietary NVLink bus for ultra-high speed communication. These appliances have ~500 TeraFLOPS of deep learning capacity each, continuously generating better algorithms to find personal data, and propose legal ground for processing.