Jerry Liu
Also known as LlamaIndex
Jerry Liu created GPT Index in November 2022, the project that became LlamaIndex. Its first problem was the limited amount of knowledge that could fit in a language model prompt. The framework added ways to load information, index embeddings, and work with vector stores. Liu’s June 2023 tutorial with Erika Cardenas explained retrieval-augmented generation through LlamaIndex and Weaviate. Liu is LlamaIndex’s co-founder and CEO, and he wrote a September 2025 post on LlamaParse. LlamaIndex’s current direction centres on document processing and workflows for agents. His public voice combines original tools with explanations of how developers can supply useful data to AI applications. The public framework and tutorials let readers follow the connections among source documents, retrieval, and the answers a model produces.
- Now at
- LlamaIndex
Timeline
-
November 2022
Starting GPT Index
Liu made his first commit to GPT Index in November 2022. The initial system addressed the small amount of knowledge that could fit into a GPT-3 prompt. It organised external information so a model-powered application could use a larger source of material.
-
Late 2022–2023
Adding ways to connect data
The project added embeddings, vector store support, and loaders for sources such as Notion, Slack, and Google Drive. The project expanded from a simple tree index into a framework for using varied data. This gave developers concrete connections between model calls and their own information.
-
June 2023
Teaching retrieval with Weaviate
Liu and Erika Cardenas published a tutorial combining LlamaIndex with Weaviate. It explained retrieval-augmented generation: search external information, then supply it to a model answering a question. The article made the architecture concrete through a data framework, a vector database, and a working setup guide.
-
November 2023
A company and open source community
LlamaIndex gained contributors and users during its first year. The LlamaIndex repository contains the implementation and documentation.
-
By September 2025
Extending into document processing
In September 2025, Liu wrote about LlamaParse and document automation a year after its public beta. He framed documents as a major source of business knowledge. The product direction linked parsing and extraction with the data required by model-powered workflows.
-
Current project direction
Building document workflows
LlamaIndex focuses on agents for document processing and workflows. Its public tools connect ingestion and extraction with later reasoning and automation. Liu has written about both the context-limit problem and preparing data that an application can use.
Sources and credits
Photo: LlamaIndex. License: Not stated by source. Source: llamaindex.ai. Cropped to a square and resized. Ask us to remove or replace this photo.