This is where Molecula positions itself as a game changer. The company closes the gap between data and decision, enabling organizations to unlock the power of realtime analytics and AI. By eliminating resource-intensive preaggregation, Molecula accelerates data-driven business innovation. “We leverage machine ready data from the original source, dramatically reducing the time, money, security, and effort required to deliver data-driven business outcomes. We help get data ready for analytics projects without all of the waste and wait to rework the current approaches adopted for every single use case,” comments Maycotte. This way, Molecula enables its customers to execute models and create value for their businesses without having to deploy dedicated data pipelines and infrastructure for every single use case.
Molecula provides its foundational technology, FeatureBase, to speed time-to-value and deliver the promise of real-time analytics and AI. By eliminating the time-intensive preaggregation, FeatureBase makes data immediately accessible, actionable, and reusable, powering real-time analytics and machine learning applications to drive instant decisions.
FeatureBase is an overlay to conventional big data systems that automatically extracts features, not data, from each underlying data source or data lake and stores them in one centralized platform. Reducing the dimensionality of the original data , FeatureBase collapses conventional schemas into a highly optimized machine-native format (10-100X smaller). This eliminates the need for incredibly expensive and time-consuming processes and allows for upto-the-millisecond data updates with little to no upfront data preparation.
Maycotte notes, “We reduce the customer’s data footprint by 10-90X,bypass the painful preaggregation step, and allow them to do their aggregations, joins, and transformations in their queries, models, codes, and applications. We can thereby shave off weeks, months, and sometimes even years from their timelines.”
We leverage machine ready data from the original source, dramatically reducing the time, money, security, and effort required to deliver datadriven business outcomes
With FeatureBase, Molecula helped one of its clients, a global leader in marketing, efficiently unravel their preaggregation complexities. The clients’ platform held around 25,000 distinct customer attributes, including millions and millions of records. As such, they were spending 36 to 48 hours preaggregating data into rollup tables that could be queried for specific ad bidding purposes. Even with the rollups, those queries were taking five minutes to return. Once Molecula automatically extracted that data into features inside FeatureBase, they were able to completely eliminate hours of preprocessing, and the same queries now return in 41 milliseconds.
In another impressive feat, Molecula addressed the challenges of an online e-banking platform serving more than 400 financial institutions across the country. Prior to using FeatureBase, the company had to think about their data at an institution-by-institution level, and it was too complex to query at more granularity. “We helped those 400 financial institutions convert their 400 databases that they had about their customers into features. This allowed them to power a real-time dynamic personalization platform that enhanced their customer experience. The platform also looks for fraud and makes analytical decisions instantaneously, and it does so with millisecond-fresh data and millisecond latencies in the products,” explains Maycotte. By analyzing the current data and allowing models to compare it to all of the historical data in the moment, Molecula can find the bad actors instantaneously, eliminating minutes, hours, or even days that go between the anomaly and the actual detection of that anomaly. Leveraging FeatureBase, the e-banking company can now add every single new use case to the platform in a matter of weeks instead of months and years. They are now running seven individual use cases off of a single FeatureBase implementation and continuously coming up with new ideas. Scripting several such success stories, Molecula is progressing in full steam and is set to launch its cloud product in the first half of 2022. “I think the companies that won the information era are also well positioned to win the intelligence era. There are hundreds of startups reinventing every industry, from agriculture to healthcare to financial services, but to survive the future, bold moves are imperative to transform corporate cultures and infrastructures to compete and stay relevant,” shares Maycotte.