Explorium Leverages AWS to Power Its Data Enrichment for ML Platform

An increasingly interconnected world has put data and Machine Learning (ML) at the center of many business decisions. Or, as the commonly repeated quote goes, “Without data you’re just another person with an opinion.”

Finding the right data, both internally and externally, for your ML can be a huge pain though. It’s often dirty, hidden behind paywalls, or just not enough to give a full view of a situation. This is where Explorium comes in.

Founded in 2017, the startup offers a platform designed to handle much of the data science process, including cleaning your internal data and managing ML models. However, Explorium’s magic really lies in the platform’s data discovery and enrichment capabilities, which allow users to automatically connect to thousands of external data sources, integrate them into their own internal datasets, and rapidly—through the platform’s own machine learning algorithms—distill the most impactful signals for the predictive challenge they are solving.

Or Tamir, co-founder and COO

“Right now, we’re in the middle of an AI/ML revolution, where companies of all sizes are looking to integrate these powerful decision-making technologies into their products,” says Or Tamir, co-founder and COO at Explorium. “The problem is that a lot of time is wasted in trying to get the right data needed to make these models accurate enough to use effectively. Explorium was founded to make that process easier, giving data science and analytics teams access to the wide potential of data sources out there, and more time to focus on projects core to the business.”

To help power it all, Explorium has partnered with AWS as a cloud provider since day one, per Tamir. “From the beginning, AWS has given us the infrastructure needed to develop our entire platform, which spans data enrichment, feature engineering, ML modeling, and running it all in production, at scale, and with high availability.”

The wide use case means Explorium leverages an array of AWS services to support its end-to-end platform. That said, the center of it all is data, which leads the startup to lean heavily on database services and compute to store, process, and manipulate the many types it manages.

“From Amazon S3 to DynamoDB and RDS, our team relies on all of them to ensure we’re able to manage countless types of data. The scalable compute options AWS offers also enables us to quickly derive relevance from the data, and then transform it, whether by automatically coming up with hundreds of thousands of ideas for different features or testing signals to see what actually helps. It couldn’t happen without AWS.”

The company also works with Amazon SageMaker, both as a user and when integrating with their own customer’s ML processes. “While some of our customers want an end-to-end experience, specifically ones that are just starting out on their AI/ML journey, we also have very advanced customers that piecemeal their own best-in-breed stack,” says Tamir. “They use Explorium to open up the data side of the process, and use SageMaker to manage the modeling part. Because of this, we are built to easily integrate with SageMaker, enabling teams to train their models on the right data, achieve better results, and then consume it in production through SageMaker and export to pipelines.”

Looking ahead, Explorium seems to be well set up for success. With a fresh $31 million Series B secured in July of 2020 and a growing industry ahead of them, the team will continue helping their customer attain and use the best data for powering robust predictive models.

Interested in ML on AWS? Contact us today!