KONUX leads the way in predictive maintenance
Utilizing machine learning for predictive maintenance is an area that holds great promise for many industries. KONUX has been a trailblazer in this domain, and they’ve been internationally recognized, including in CBInsight’s AI 100 in 2020 and 2021. The company’s technology is utilized by some of the largest railway companies in the world, including the railways in Germany, the United Kingdom and China.
Combining IIoT and machine learning
KONUX combines IIoT and machine learning to transform railway operations. Their platform allows railway operators to monitor their rail infrastructure and inform them about maintenance issues before they happen. The company’s technology is utilized by some of the largest railway companies in the world, including the railways in Germany, the United Kingdom and China.
Their IIoT devices are placed along the track, measure vibration from passing trains. In addition, KONUX has built intelligent analytics that can detect and identify trains and railcars on the rail from the sensor data and alert to any abnormalities. This helps the railway operators dispatch maintenance crews to ensure safety, improve availability and prolong the lifespan of the existing infrastructure.
Utilizing machine learning for predictive maintenance is an area that holds great promise for many industries. KONUX has been a trailblazer in this domain, and they’ve been internationally recognized, including in CBInsight’s AI 100 in 2020 and 2021.They are also a World Economic Forum AI Council member.
From moonshots and day-to-day operations
KONUX chose Valohai to be their MLOps platform in 2019. The data science team uses Valohai to continuously train production models that provide insights within their SaaS product. In addition to that, they also run exploratory research on Valohai. The team calls such experimental projects moonshots. Each individual can spend 15% of their time on moonshots to find new breakthroughs.
At KONUX, the best solutions are found through rigorous, large-scale experimentation. Valohai maintains records of all experiments and orchestrates the required machine resources.
The team has seen significant productivity gains using Valohai because they can launch multiple simultaneous experiments and return to the results later. Each experiment may take from minutes to days to complete, so it’s paramount that data scientists don’t have to monitor the infrastructure.
Previously, a data scientist would manually spin up machines to run an experiment and then wait to shut them down to avoid accruing unnecessary costs. In an environment where innovation and experimentation are encouraged, workflows have to support that. Valohai’s automatic machine orchestration and experiment tracking have enabled the team to 10X the number of experiments run with the same amount of effort.
Large-scale experimentation tends to be tricky because you’ll need to manage cloud resources, and mistakes can be quite costly. With Valohai, though, that stress is gone, and we can focus on the actual data science. The version control of all parts of an experiment, from code to data to environment, allows for systematic research, which can be reviewed months later. On top of that, when we do all of our experiments on Valohai, it’s easy to promote them for production use later on.Andres Hernandez – Lead Data Scientist, KONUX
The moonshots are presented at a company-wide event called "The Festival of Ideas". The successful experiments that work out both in technical and business terms often make the most important new product features. With all experiments being stored on Valohai, it’s easy for the team to forward the best ones to production.
Building the future together
The experts at KONUX are avid users of Spark to crunch the massive amounts of sensor data. Together with them, we’ve developed Valohai to support Spark clusters (starting with AWS EMR). We’re continuing the development to raise Spark to be a first-class citizen in the platform with feature parity to all other executions. The team at KONUX has helped us prioritize capabilities that help them get their job done.
The quick response from the Valohai team to new requirements and use cases is an important reason for KONUX’ choice of Valohai. The need to connect data extraction and preprocessing steps that run in Spark into a Valohai pipeline went from the first discussion to a prototype in less than two months, and by three months, it was feature complete.
For us, Valohai isn’t just another tool; they are a partner in supporting our growth. Valohai continues to support us with new product features that help take machine learning further in our field.Olav Stetter – Head of Data Science, KONUX
To hear more about how KONUX utilizes Valohai, reach out to us.
To read more about KONUX, visit www.konux.com
The best of the best
MLOps in the WildA collection of MLOps case studies
The MLOps space is still in its infancy and how solutions are applied varies case by case. We felt that we could help by providing examples of how companies are working with tooling to propel their machine learning capabilities.
Think of this as a lookbook for machine learning systems. You might find something that clicks and opens up exciting new avenues to organize your work – or even build entirely new types of products.Download