Case Study: Festo reduces machine downtime and speeds ML experiments with Weights & Biases

Festo cuts experiment setup time from 8 hours to 20–30 minutes with Weights & Biases

Festo, a global leader in industrial automation, faced the challenge of preventing costly machine downtime for its customers. To move from routine preventive maintenance to predictive analytics, its ML team needed a more efficient way to develop and tune models that could identify equipment failures, such as air leaks in pneumatic cylinders, before they occur. They partnered with Weights & Biases to streamline this process.

By implementing Weights & Biases' MLOps platform, including Sweeps for hyperparameter optimization and Launch for scalable training, Festo drastically improved its workflow. The solution reduced experiment setup time from an average of eight hours to just twenty to thirty minutes. This allowed the team to focus on iteration and dramatically scale up model training, ultimately helping to transform traditional factories into smart factories with less downtime.

Festo