The average yearly AI expenditure for most companies in the United States will be estimated to hit about $14 million this year alone; this could prove to be unsustainable for smaller to medium companies who may lack the senior expertise that supplants the reliance for this AI tools. Sherlock aims to solve this cost problem by taking a novel approach to inference to reduce infrastructure costs. Sherlock is also built with security in mind, it has been designed to work with private clouds, off-prem hardware and locally hosted LLMs.
