Article
eomer reached the semi-final of the NVIDIA Grand Inception Challenge, which brought together startups from across Southeast Asia. During the competition, the team presented the eomer platform and its approach to structured-data decisions to the panel.
The platform addresses prediction tasks that rely on tables, time series, and enterprise records. Relevant use cases include energy-price forecasts, retail-demand forecasts, vessel-arrival estimates, advertising response, and other regression or classification problems. These use cases differ by sector, but each requires a model that teams can test against historical outcomes before deployment.
The eomer approach uses tabular foundation models through a common workflow. A team can prepare a dataset, establish a baseline, run inference, inspect uncertainty, and compare results against defined evaluation criteria. This process reduces the need to create a separate model-development workflow for every new dataset.
The competition also provided an external setting in which to explain the product to an audience that had not participated in its development. That setting tested whether the team could communicate the problem, the technical approach, and the commercial application within a limited presentation format.
Reaching the semi-final provides one point of external validation, but product performance remains the relevant deployment criterion. The next steps therefore remain focused on customer datasets, historical backtests, production integrations, and measurable improvements within specific decision processes.
