Case Study: Constru boosts experiment reproducibility and AI productivity with ClearML

Constru cuts experiment reproduction time by 50% with ClearML

Constru, a company using advanced computer vision to analyze and control the construction process, needed to streamline its AI experiment management workflow. The challenge was to rapidly extend its platform's capabilities without adding staff, which required its researchers to spend less time debugging code and managing experiment documentation and more time on core research. They sought a solution to ensure experiment reproducibility and avoid errors that could lead to wasted time or incorrect conclusions.

The vendor ClearML provided its open-source MLOps platform, which automated experiment documentation, facilitated team collaboration, and simplified the process of reproducing and comparing experiments. This solution enabled Constru to run up to seven times more concurrent experiments. The measurable impact included an estimated 50% reduction in the time needed to reproduce experiments and projected savings of $1.3M over 12 months, allowing the company to significantly improve its product feature release rate without hiring additional staff.

Constru