LlamaIndex Raises $8.5M to Unlock Large Language Models Capabilities with Personal Data

LlamaIndex, the data framework for Large Language Models (LLMs), announced it raised $8.5M in seed funding led by Greylock. Angel investors in the round include Jack Altman, Lenny Rachitsky, Mathilde Collin (CEO of Front), Raquel Urtasun (CEO of Waabi), Joey Gonzalez (Berkeley) among others. The company will use the funds to build an enterprise offering on top of the popular open source project with the same name.

LLMs have a huge number of potential use cases, yet they are limited to the data they are trained on. While exploring the possibilities of GPT-3 this fall, ML researcher and engineer Jerry Liu discovered limitations specifically around the models’ ability to work with data private to an individual or enterprise, from files (PDFs, Powerpoints), to workplace apps (Notion, Slack, Salesforce), to databases (Postgres, MongoDB).

“Many users building applications on top of LLMs want to unlock new use cases with their own private data,” said Jerry Liu, CEO & Co-founder of LlamaIndex. “To solve for this, I created an open-source project called LlamaIndex to help unlock the full capabilities and use cases of any LLM for both myself and other developers.”

Marketing Technology News: Why a Style Guide is Pivotal to Brand Identity

The open source project quickly resonated with the AI community, from hackers to developers at startups and big companies alike. After just six months, the project has 15K Github Stars, 19K Twitter followers, 200K Monthly downloads, and 6K Discord users.

Seeing teams at companies like Uber, Instabase, and Front use LlamaIndex to prototype LLM-powered features over their data, Jerry recognized the project could be incredibly helpful to enterprises, who often want to ingest and structure their private data in a way that can be used with LLMs, and perform various LLM tasks over that data and obtain high-quality responses with verifiable sources. In March, Jerry joined forces with Simon Suo (Co-founder and CTO) who he met while working on ML research at Uber, to start LlamaIndex.

Today, LlamaIndex’s open source toolkit can be used with any LLM and offers an easy to use and optimized experience. The toolkit can handle a diverse range of data sources, from structured to semi-structured to unstructured text or even image data, and includes data ingestion, data indexing, and a query engine layer on top. The company’s enterprise solution – which will be available later this year – will be built on top of the popular open source project. The solution will help enterprises eliminate the technical and security barriers for data usage, and provide a range of services, from scalable/reliable data source connectors to security features such as access control and user management.

“LlamaIndex’s early traction within the open source community is a reflection of the project’s simplicity and short time-to-value.” said Jerry Chen, Partner at Greylock. “Jerry and Simon are strong AI technologists who have worked across both research and engineering at companies like Uber, Robust Intelligence, and Quora, and are exactly the type of founders we love to back.

Marketing Technology News: MarTech Interview with Carol Howley, CMO at Exclaimer