Michael Berthold, Founder and CEO of KNIME, shares insights on the company's evolution, addressing current data science challenges, and empowering developers.
In the rapidly evolving world of data science, companies are constantly seeking tools and platforms that can help them harness the power of their data. KNIME, an open-source data science platform, has been at the forefront of this revolution, providing a comprehensive environment for data preparation, machine learning, and analysis. I recently had the opportunity to catch up with Michael Berthold, Founder and CEO of KNIME, at the Snowflake Data Cloud Summit, where we discussed the company's journey over the past five years and its vision for empowering developers, engineers, and architects in the data science landscape.
Evolving With the Times
Over the past five years, KNIME has undergone significant changes to stay ahead of the curve. "We completely changed both of our technologies," Berthold revealed. The analytics platform is now browser-ready, and the KNIME server has been replaced with a cloud-native business hub. The company also recently launched a SaaS offering, allowing users to access KNIME's powerful features without the need for on-premises installation.
When asked about the recent announcements made at the Snowflake Summit, particularly regarding Cortex and Polaris, Berthold expressed his thoughts on their potential impact. "With Cortex AI, they are following up on Databricks," he noted, adding that the rapid three-month launch timeline was surprising from a software engineering perspective. However, he acknowledged the importance of AI integrations and emphasized that governance is a huge topic, particularly in the realm of data science.
Staying Ahead of the Competition
KNIME's role in the data analytics and integration ecosystem continues to evolve, and the company remains focused on allowing organizations to leverage their data science IP effectively. Berthold highlighted the significance of the IP involved in building workflows, stating, "When you change your execution platform, or your data storage platform, or your back end, something else, whatever comes up to Excel for visualizations, or from Power BI to Tableau. The core IP stays as part of a KNIME workflow, and you only adjust the connectivity."
KNIME's partnerships and collaborations play a crucial role in its growth and competitiveness. Berthold mentioned the company's recent two-year collaboration with Harvard, focusing on geospatial analytics. This collaboration has resulted in a tight integration between KNIME and Harvard's geospatial analytics capabilities.
Addressing the Challenges of Data Science
One of the most significant challenges faced by companies in the data science domain is the flood of tools available. Berthold pointed out that while many organizations create fantastic reports, insights, and predictive models, moving those into production remains a problem. "Then you buy something else from another vendor, but it's not quite compatible. But your move into production isn't quite what you have in mind," he explained. KNIME aims to address this challenge by providing a platform that seamlessly integrates with various tools and enables smooth deployment of data science workflows.
Berthold also highlighted the importance of using the right technology to avoid the need for feature stores. "If people had used the right technology, they wouldn't need to use feature stores because feature stores fundamentally try to fix the problem of most models not containing feature transformations," he stated. KNIME's approach captures both the model and the necessary feature transformations, eliminating the need for separate storage facilities.
Empowering Developers With Low-Code and Visual Programming
KNIME's focus on low-code and visual programming sets it apart from other data science platforms. Berthold clarified that there are two types of low-code models: those that use programming languages under the hood and translate visual representations into code, and those like KNIME, where the workflow itself is a visual representation. "It's a visual way of putting together what you want to do with your data. And underneath, it does call out to all sorts of Python, SQL, other pieces of code, other libraries, but you don't necessarily need to interact with those," he explained.
This visual programming approach empowers developers to build complex data science workflows without getting bogged down in coding details. KNIME's extensive plugin ecosystem and integration capabilities further enhance its flexibility and adaptability to various data science projects.
Looking Ahead
As KNIME continues to evolve, the company remains committed to empowering developers, engineers, and architects in the data science realm. With a strong focus on governance, AI integration, and ease of use, KNIME is well-positioned to address the challenges and opportunities that lie ahead.
Berthold expressed his excitement about the upcoming release, which will introduce a new user interface and improvements to workflow interaction. These enhancements aim to streamline the data science process and make it even more accessible to a broader range of users.
In a world where data is increasingly crucial to business success, platforms like KNIME play a vital role in enabling organizations to extract valuable insights and drive innovation. As Michael Berthold and his team continue to push the boundaries of what's possible with data science, developers, engineers, and architects can look forward to a future where they are empowered to tackle even the most complex data challenges with ease.
Comments