Blog | 2 weeks ago | 4 — 6 mins

Today, I’m excited to announce the next chapter in Kaskada’s journey: joining DataStax to deliver a powerful, unified solution for real-time AI. And while there will be lots of formal announcements over the coming weeks and months, I wanted to share a few words of my own.

When we founded Kaskada five years ago, we had a singular focus to massively expand the reach of data science and machine learning, stripping out the complexity and cost of feature engineering and feature serving. This is among the most important, most painful work for ML practitioners. While the industry has given significant attention to model serving, feature engineering and feature serving have been largely neglected, addressed instead through massive headcount investments, the creation of new cost centers, and with it, spiraling costs for solutions that still fell short.

This early work helped us understand the value of the right abstractions, connecting directly to raw event-based data, helping to accelerate iteration, and bridging the gap to production— an effort that finally made behavioral ML feasible.

Fast forward to 2023 and today’s news. We are absolutely thrilled to join a team with a shared vision for delivering great technology that supports the growing speed and scale requirements of businesses today, along with a commitment to open source and co-innovation with a global community. Even more, DataStax holds a unique connection to Apache Cassandra, the massively scalable database used by more than 90% of the Fortune 500. AI is at its best when it has access to data at scale. And real-time data, in particular, will shape a generation of new applications and real-time decisions for every industry. Cassandra is uniquely suited for vast amounts of real-time data, making our decision to join DataStax strategically relevant for us, our mission and the market.

Together we can deliver a new generation of AI-powered applications, reshaping the relationship of ML and data, bringing ML to the real-time data already serviced by Astra DB and Cassandra. We can eliminate the cost, time and effort of transferring data to discrete ML systems. And we can give every business access to the real-time data that enables them to make more accurate and more powerful decisions at the exact time they are needed. In short, together, we will make real-time AI a reality for all businesses.

As I reflect on our journey at Kaskada, I’m very proud of what our team has accomplished. To achieve this success, especially at such an unusual time for the industry, required the perfect combination of people, timing and technology.

How we got here: technology, people, timing

At the core of Kaskada’s success is the amazing technology that we’ve developed: processing​ event-based data​ to train​ behavioral ML ​​models ​and delivering ​instant, actionable experiences that fuel real-time customer interactions in production​. With Kaskada’s patented technology, practitioners directly connect to event data generated by real-time applications to get more value from the data to power predictions, decisions, and experiences. Kaskada’s feature engine brought iterative time-based feature engineering that automatically protects from leakage with native time-travel and data-dependent windowing and provides a robust data infrastructure for computing, storing, and serving features in production.

The talented people who built the technology and took the company into its API-first form showed incredible tenacity and resilience. Over the years, we’ve had people join, leave, endure a pandemic and succeed during the toughest year for tech in two decades. We’ve lived through many moments together, from celebrations to disappointments, and every single one of these people were able to deeply contribute to this success despite the environment around us.

Finally, the timing. Our vision back in 2018 began ahead of its time. For years, the talk around real-time ML was mostly hype. Many companies we spoke to implementing ML on the ground still needed to spend years testing the limitations of the batch-oriented systems first before considering a proprietary offering in the real-time space. But, as time passed, many realized that the limitations were too great of a trade-off to accept. The majority of these initiatives failed to make an impact on the business at scale instead being relegated to analytics dashboards across a subset of data.

Those who had the budget began to convert their systems into hybrid batch and streaming systems. Attempting to gain the benefits of real-time ML for users and the known patterns of batch for iteration and discovery. Just recently, industry leaders in this space have concluded that real-time ML is the way to go, something that required 5-7 year journeys from batch to real-time ML for leaders like Netflix and Instacart.

At the same time, the fit between DataStax and Kaskada became too good to pass up. While we didn’t set out for Kasada to be acquired, the opportunity to achieve more, together, and faster, is a great one for Kaskada, our customers, and the market.

What’s next: Accelerating our vision with DataStax

In joining DataStax, we bring together leaders in two fields with complementary strengths, missions and visions. Our core efforts continue with even more conviction, to advance data science and machine learning, to significantly simplify feature engineering and feature serving and to deliver the most accurate, most powerful insights into customers, operations and markets through real-time. Even more, we will achieve this in partnership with the open source community to deliver an open, unified solution for real-time AI.

Together, we open up the possibility to change the world. It is an incredible opportunity and honor to help DataStax make real-time AI a reality. I look forward to the years ahead and continuing to serve this amazing community.

To learn more about the power of real-time AI, check out the white paper and video and sign up for ongoing updates, product details, and additional insights into how DataStax is making real-time AI a reality.