Join top executives in San Francisco July 11-12 to hear how they are integrating and optimizing their AI investments for success. learn more
ClearML co-founder and CEO Moses Guttmann knew the potential of his company’s open source-based MLOps tool. What he didn’t know was that in late 2022, when ChatGPT came out, the whole market and his company would accelerate.
Today, ClearML announced a series of platform updates, with over 1,300 global enterprises now using the ClearML MLOps platform, along with strong growth in Q1 2023. The growing demand and interest in developing and deploying machine learning (ML) models is driving momentum, and organizations of all sizes are poised to benefit from this technology.
The basic idea of MLOps is to give organizations the tools they need to manage their machine learning build and test workflows. ClearML has both an open source project and an enterprise edition debuting in September 2022.
Among the new features ClearML is launching is a feature the company calls “Sneak Peek” that goes a little beyond the traditional MLOps feature. Sneak Peek allows users to iteratively deploy and preview a test model in real time while the model is still under development. ClearML also adds new model lineage capabilities to help with AI explainability.
event
transform 2023
Join us July 11-12 in San Francisco. A top executive shares how she integrated and optimized her AI investments and avoided common pitfalls for success.
Register now
Guttmann told VentureBeat: “We believe there is a lot of interest in the ChatGPT hype, and we basically understand that everyone really has to be on board.”
Sneak into the future of model development
An MLOps workflow typically includes a series of steps that help a data scientist build a model.
What ClearML does with its sneak peek approach is to make it easier for data scientists to deploy internal machine learning-based applications for products and business units to experience as part of their development process. According to Guttmann, the goal is to make his ML development more accessible and reduce the time it takes organizations to get value from the overall process.
“Before this, ClearML was aimed at an audience of machine learning engineers and developers,” Guttmann said. “At Sneak Peek, we also target product people.”
An emerging use case that Guttmann sees is implementing ML directly within the product using a continuous learning approach. He pointed out that there are organizations using ClearML, where models are constantly being trained as data is collected.
“We have seen companies deploy machine learning automation as part of the product itself,” he said. “So the product itself has this training capability.”
Improving AI Explainability with Model Lineage
Another area of improvement in ClearML is the addition of new model lineage capabilities.
Model lineage allows organizations to track where different elements of a model come from and how they change over time.
“Over time, being able to run some forensics on deployed models becomes very important,” says Guttmann. “So if something goes wrong, we can trace the original codebase and data that was used to train that particular model.”
With model lineage, ClearML now provides clear visualizations to help understand who created the models and where they are being used in production, he said. Being able to track lineage is a key component of AI explainability and helps organizations empirically track what has been done in model development.
“I try to advocate for safe and secure model development,” he said.
Mission of VentureBeat will become a digital town square for technical decision makers to gain knowledge on innovative enterprise technology and trade. Watch the briefing.
