AIEngineering
#aistrategy #gpu #infrastructure #tpu
Most of the AI Strategy today focuses on data, tools and people. One key component everyone miss out is for accounting the right infrastructure and hardware accelerators to accelerate AI initiatives
Data scientist typically have to go through 100’s of training iteration for each hypothesis to narrow down on feature and model selection, hyper parameter configuration among others. Data in many cases are way beyond what a single node or GPU can fit in. For large datasets each iteration can go upward of days. It is very important to plan for Infrastructure either cloud on on-premise that can reduce frustration among data scientist as well as help enterprise deliver enterprise faster
Source
