Adaption's AutoScientist Co-Optimizes Data and Model Training Simultaneously

Adaption unveiled AutoScientist, a self-training tool that optimizes training data and model parameters in tandem rather than sequentially. The platform collapses the traditional two-phase workflow into one, claiming more than doubled performance metrics across various models. It builds on Adaption's Adaptive Data product and addresses a persistent bottleneck: the iterative experimentation required for production-ready AI. Sara Hooker, former VP of AI research at Cohere, co-founded the company.