Predibase offers a low-code, declarative ML platform aimed at developers to build, train and deploy models with minimal code. The platform exited stealth last year and during its beta users trained over 250 models. Predibase now supports deploying custom LLMs and provides its own LudwigGPT LLM, plus a Data Science Copilot that recommends model improvements. The company has a free two-week trial and lists customers including Paradigm and Koble.ai. Predibase says it will use the new funding to expand go-to-market efforts and continue building out the platform amid competition from large cloud vendors and other startups. Predibase emerged from stealth with a platform that lets users specify an AI system as a file and uses a declarative approach to build models. The product is built on open-source technologies including Horovod and Ludwig and is designed to define AI pipelines in a few lines of code while scaling to petabytes of data across thousands of machines. Predibase integrates with data sources such as Snowflake, Google BigQuery, and Amazon S3 and supports modalities including images, text, tabular data, and audio, with model explainability built in. The company runs an invite-only beta and reports usage at Fortune 500 customers, including a leading U.S. tech company, a large national bank, and a large U.S. healthcare company; its team totals 21 people. Predibase raised $16.25 million in a Series A and plans to use the capital to expand the beta to a wider market, grow its machine-learning engineering team, and build out a go-to-market organization. Co-founders include Piero Molino and Travis Addair, and early team members/co-founders include Devvret Rishi and Chris Ré.

Total raised

$28M

Funding rounds

2

Latest round

Series A

Latest activity

May 2023