Sergey Nivens - Fotolia

Airbyte set to advance open source data integration platform

The CEO of a data integration startup explains how he is looking to grow the technology and community behind the vendor's open source data extract and loading platform.

Data integration startup Airbyte said on March 2 that it raised $5.2 million in a seed round of funding led by Accel to help the vendor build out its platform and accelerate go-to-market efforts.

Airbyte has developed an open source ELT (extract, load, transform) platform with code available on GitHub that enables anyone to use the technology for free. The company also has a cloud-hosted offering that provides Airbyte as a managed service.

Getting data from one source into another often involves the use of ELT technology.

ELT helps enable data integration, in which data can be loaded from one source and then used for business intelligence, data analytics or inside of a data pipeline to enable an application. The market for data integration is competitive, with multiple vendors in the space, including Fivetran and Talend, among others

Airbyte, based in San Francisco, is a new entrant to the market.

In this Q&A, Michel Tricot, co-founder and CEO of Airbyte, discusses the challenges and opportunities of enterprise data integration and where his company is headed.

Headshot of Airbyte CEOMichel Tricot

Why are you now raising money for Airbyte's data integration efforts?

Michel Tricot: I've always worked on data since I started my career and I've seen firsthand the problems related to data integration. My co-founder John Lafleur and I got started spending a few months in late 2019 brainstorming ideas. In 2020, we went through startup accelerator Y Combinator and interviewed people in as many companies as we could to understand their data integration challenges. The problem we want to solve is to enable people working on data to be able to focus on extracting insight instead of extracting the data.

Airbyte was first released at the end of September 2020 and we've got great traction for it so far. The fact that we're getting that traction shows that somehow we are in a blue ocean with our approach, and that people are actually waiting for this approach. If you're in a blue ocean, you want to grow as quickly as possible and that's why we decided to do the fundraise now.

The data integration space is not new. What do you hope to bring to the table that isn't currently in the market?

Tricot: We talked to a lot of companies that were already using data integration tools and it was quite a surprise to see that although they were paying for these tools, they were still building their own tooling to actually enable the data integration. There are a few reasons for that. First, there is a continuously growing number of sources where data is siloed, meaning that every single company, if they want to support new sources, ends up building new data connectors themselves.

The other use case is customization. Often data integration technology is designed with a specific use case in mind. That's why we built Airbyte as open source, so anyone can contribute to the platform. Open source also allows our users to play around with these connectors and to actually customize them for their own needs.

The problem we want to solve is to enable people working on data to be able to focus on extracting insight instead of extracting the data.
Michel TricotCo-founder and CEO, Airbyte

How does Airbyte actually load data to enable data integration?

Tricot: Today there is no standard for exchanging data. When you're thinking about change data capture, generally that's a technology that is limited to database. With Graph QL every data source you connect with has to support GraphQL. With REST APIs, everybody has a different way of setting up REST APIs.

What we are doing, is standardizing the data access, taking into account the specificity of the different systems that are being connected. We try to stay away as much as possible from the transformation side of data integration and we really focus on extracting and loading. The only transformation that we do is ensuring that the data can land in the destination.

So, if it's a data warehouse, it means we're going to make sure that the data looks like a table in a database. That's the limit of the transformation we're doing, we want to stay as close as possible to what the data looked like in the origin, so that we don't apply any bias to the data.

What's next for Airbyte?

Tricot: Goal No. 1 is we need to continue to grow the community, so that's our main focus and that's where we're going to dedicate a lot of energy.

Now, to gain that community we also need to continue to deliver on our promises and deliver new features. So we'll be adding new connector over the course of the year and we also want to support Kubernetes with a nice integration.

It's really about continuing to build the platform and the community.

This interview has been edited for clarity and conciseness.

Next Steps

Fivetran looks to Teleport Data for database replication

Dig Deeper on Data integration

Business Analytics
Content Management