Case Study: Lensor Achieves Faster, More Reproducible Computer Vision Development with ClearML

Lensor runs dozens of experiments in parallel with ClearML

Lensor, a technology company focused on AI-driven vehicle inspection systems for the automotive sector, faced significant MLOps challenges as it scaled. Their team struggled with fragmented management of hybrid compute resources (on-premises GPUs and AWS), a manual and inefficient model lifecycle, and a lack of data traceability and centralized experiment tracking. They needed a unified platform to streamline their entire AI development process.

By implementing ClearML's open-source platform, Lensor automated its entire workflow. The solution utilized ClearML Agents for seamless orchestration of hybrid compute, HyperDatasets for versioning and managing large-scale image data on AWS S3, and Pipelines to automate data and training lifecycles. This resulted in faster development cycles, more robust models, and the ability to efficiently scale operations. The ClearML implementation provided rigorous experiment tracking for full reproducibility and eliminated previous bottlenecks, empowering the team to innovate faster.

Lensor