Rockset and the Future of Data-Driven Apps

One of the most important technology themes of the past 10 years has simply been data.

At the beginning of the decade, new data techniques helped drive an increased focus on analytics and A/B testing. In the past few years, the focus has shifted to machine learning. Both technologies are rooted in the same trend — the exponential increase in data.

The CEOs we meet know that their businesses need to become data-driven to compete and survive. But not everyone knows how.

Getting started is easy — you instrument your product or business and start collecting data. But turning that data into valuable, actionable insights is much more challenging.

One of the biggest pain points is just physically moving the data around. The data pipelines and ETL jobs that move data from its initial raw form to the final product are often slow, fragile, and hard to maintain.

There have been many efforts to simplify these pipelines, but the state of the art is still Rube-Goldberg-esque. It can take weeks or months for new types of data to make their way into production applications — if they ever make it at all.

Rockset

At Sequoia, we love partnering with founders that challenge conventional thinking.

When we first met Venkat and Dhruba, they wondered — why do we need data pipelines at all? In this new era of cloud-native infrastructure, why can’t we use web-scale search techniques to build data-driven applications directly on top of the raw data? Why make pipelines 10% better when we can eliminate them entirely?

Their answer is Rockset. A cloud-native service that helps developers run production-ready SQL directly on top of raw data. It’s super simple to get started:

  1. Create an account
  2. Point Rockset at your data source (e.g., Apache Kafka, Amazon S3)
  3. Start writing SQL

Once you’ve configured your data source, Rockset immediately starts indexing your data, enabling you to explore your data and write application-ready SQL within seconds. At Sequoia, we’ve used Rockset to move many cumbersome nightly jobs to real-time dashboards. Not only are our internal tools faster, but our data science team now spends less time babysitting pipelines and more time helping our companies grow.

Rockset is the result of a wonderful, interdisciplinary data team coming out of Facebook. CEO Venkat Venkataramani was the founder of TAO (Facebook’s online graph database) and a database engineer at Oracle. CTO Dhruba Borthakur was one of the key architects of Facebook’s data warehouse and the co-creator of both RocksDB and HDFS. Tudor Bosman was a co-creator of Unicorn (Facebook’s internal search backend) and the Gmail backend. Shruti Bhat held senior product roles at both VMware and Oracle.

We first partnered with Rockset at the seed stage in 2016, along with our friends at Greylock. At the time, Rockset was a rough idea inspired by the idea of converging a traditional SQL database with a cloud-native search engine. Today we’re proud to deepen that partnership with Rockset’s Series A financing.

It’s been a fantastic journey with the team so far — and it’s just getting started. We’re incredibly excited to see what you build with Rockset.