GS Caltex automates Canvas-to-Pipelines handoff, cuts deployment to hours

GS Caltex, a South Korean energy and chemical company, uses a division of labor between process engineers and ML engineers. Process engineers build and validate models in SageMaker Canvas without writing code, drawing on their process knowledge. ML engineers configure SageMaker Pipelines workflows for deployment and scheduled inference. The integration work required to move assets from Canvas to Pipelines was previously a bottleneck, but GS Caltex automated this handoff while reusing the assets created in SageMaker Canvas. As a result, ML engineers now deploy models within a few hours, compared with about a week before. The company also runs batch inference every 30 minutes, which reduces the need for continuously running endpoints and, according to GS Caltex, has meaningfully lowered inference costs. Process engineers began proposing applications in other processes after seeing their models used in production, and ML engineers spent less time repeating integration work and could focus on operating and monitoring the pipelines. The solution reuses validated preprocessing flows and models, and GS Caltex reports that this approach has reduced repeated work and enabled faster deployments.