Deepchecks, a machine learning validation company, has unveiled its open-source solution for continuous validation of machine learning models. The company also announced $14 million in seed funding, with Alpha Wave Ventures leading the investment round, along with participation from Hetz Ventures and Grove Ventures.
While machine learning is rapidly gaining traction and creating significant value, many machine learning models fail to make it to production or encounter challenges that result in exceeding time and budget constraints or spectacular failures. Deepchecks aims to address these issues by introducing testing and validation practices inspired by the lessons learned from software development.
Deepchecks enables practitioners, developers, and stakeholders to go beyond traditional MLOps (Machine Learning Operations) and gain visibility and confidence throughout the entire machine learning lifecycle, from development to deployment and operation in production. By incorporating thorough testing and validation at each step, Deepchecks improves the quality and reliability of machine learning models.
Philip Tannor, co-founder, and CEO of Deepchecks, highlights the importance of systematic testing and validation in machine learning, emphasizing that deploying and monitoring untested software is not a sensible approach. Deepchecks’ community-driven MLOps framework provides a clear understanding of how machine learning applications perform from research to production.
The founders of Deepchecks, Philip Tannor and Shir Chorev, have been working together since their early years and have extensive experience in machine learning, gained through programs such as the IDF’s Talpiot and the elite 8200 intelligence unit. They initially started Deepchecks as a traditional MLOps solution but realized the need for an open-source offering to cater to data scientists and developers on a larger scale.
Deepchecks offers an open-source solution that allows users to customize and reuse components to comprehensively test machine learning models and datasets. These components have been developed based on years of collective expertise and have been successfully deployed in production environments.
The solution includes monitoring capabilities, root cause analysis for production environments, and a comprehensive user interface. Deepchecks has garnered over 500,000 downloads and is currently utilized by companies such as AWS, Booking.com, Wix, as well as in regulated industries like finance and healthcare. The enterprise version of Deepchecks provides additional collaboration and security features.
Yuval Rozio, Director of Alpha Wave Ventures, emphasizes the lack of systematization for quality assurance in machine learning and the need to address this gap. Deepchecks aims to advance the field of machine learning and help businesses by providing systematic and reliable quality assurance practices.
With its open-source solution and focus on validation throughout the machine learning lifecycle, Deepchecks is poised to contribute to the growth and improvement of machine learning models, enabling organizations to deploy and operate them with greater confidence and effectiveness.
