ZOA

ZOA is a quantitative forecasting model builder that applies large-model approaches to forecasting problems. It addresses the domain-specific nature of traditional quantitative models, where practitioners spend significant time testing features, tuning hyperparameters, and iterating architectures within narrow domains. The site states that large models can find patterns that people and specialized models could not, and that forecasting generalizes across contexts.

The platform embeds large language models in optimization loops evaluated with quantitative metrics, automating the build-test-improve modeling cycle. Its models leverage data from across contexts and rely less on human intuition than existing forecasting models. The site states that, for now, the models trade. ZOA is presented as a research-oriented product, with the site inviting applications for an ML Research Engineer role and providing contact information for inquiries.