Python is the programming language of choice for artificial intelligence, however, most organizations are unable to incorporate their Python developers into legacy data infrastructures. This means they miss out on the opportunity to bring the benefits of AI into their organizations. However, developers designing around AI workflows have been lacking in open source Python libraries.
DLT (short for Data Load Tool)A Berlin-based startup thinks it may have a solution. It’s building an open-source Python library it says is designed for a new wave of artificial intelligence.
The startup says its library will integrate into existing workflows, including Python workflows that have not previously loaded data, such as Google Colab notebooks, AWS Lambda functions, Airflow DAGs, or the GPT-4 auxiliary documentation or the GPT-4 development playground.
The startup has now raised $1.5 million in pre-seed funding Mining Venture Capital Firms, founded by Ross Mason, who created the Mule project and founded MuleSoft (MULE:NYSE). Also joining the round are artificial intelligence and business founders at Huggingface, Instana, Miro, and Matillion, among others.
CEO Matthaus Krzykowski told me via email: “Most GPT-4 applications shown in the press are demo software. Users who try it quickly give up. Other AI tools (vector databases/frameworks) that venture capitalists have invested heavily in recently face many similar challenges.”
He says dlt now has a growing community of Python developers and is “deploying in production at several tech companies at scale,” including Harness, a San Francisco-based software delivery company we’ve previously reported on.
Alexander Butler, senior data engineer at Harness, said in a statement: “Leveraging dlt has transformed our data operations. It … has accelerated the pace of our DataOps team: we spend less time on EL (extract and load) and more time on T, while still being able to deeply customize our extractors as business needs evolve.”
Julien Chaumond, CTO/co-founder of Huggingface and angel investor at dltHub added: “The current machine learning revolution has been fueled by a Cambrian explosion of open-source tools in Python that have become so accessible that a wide range of practitioners have access to them. As an easy-to-use Python library, dlt is the first tool available to this new crowd.”
Regarding potential competitors, Krzykowski acknowledged that these startups “compete with ETL companies like Meltano, Stitch Data, Airbyte, and to a lesser extent Fivetran.”
However, he said, “In a bigger picture, we operate in the space of data warehouse companies like Snowflake, Databricks, Microsoft Fabric, etc. that are also looking to bring AI into the enterprise.”