Stephen Chin

Stephen Chin is Neo4j’s VP of Developer Relations, identified in that role in an August 2026 interview, and a presenter and coauthor on GraphRAG—generative AI systems that use knowledge graphs to store and retrieve information for language models. His public argument centers on the need for structured knowledge when models work with enterprise-scale information. He uses the handle @steveonjava.[9][5][3]

In his August 26, 2026 interview with Alan Shimel, Chin argued that strong model performance on small datasets does not establish readiness for enterprise workloads. Techstrong describes his example of a home-lab digital twin using pgvector and GraphRAG: an agent retrieves a relevant slice of graph information and supplies that context to the model. The example illustrates his emphasis on selecting useful knowledge before asking a model to reason over it.[9]

At the 2025 AI Engineer World’s Fair, Chin and Neo4j helped host the GraphRAG track. On June 4, he presented “Agentic GraphRAG: AI’s Logical Edge” and joined Michael Hunger and Jesús Barrasa for “Practical GraphRAG - Making LLMS Smarter with Knowledge Graphs.” His Agentic GraphRAG presentation explains vector search with graph context as using graph patterns to retrieve items related to the initial vector-search results.[1][2][3]

Chin, Hunger, and Barrasa are also credited together on GraphRAG: The Definitive Guide, which O’Reilly lists for January 2027. Their public companion repository describes a fictional consultancy staffing projects from CVs, project reports, and HR data. The worked examples progress from a basic vector-search assistant to graph-backed retrieval, agentic tool calling and MCP, persistent memory, and evaluation of agents and knowledge-graph construction.[4][5]

In his July 2025 DevRelCon talk, Chin proposed an Engaged Developer Funnel in place of judging developer relations primarily by sales leads. His framework asks whether developers know the basics, understand when to use—and when not to use—the technology, and keep it top of mind. As he put it, “And if we teach them when not to use it, they won't misapply it and they won't recommend other people use it for projects where they can't be successful.”[6]

Chin described applying the framework at Neo4j through training-course participation, cloud query activity, and continued contact through newsletters and community forums. He said the metrics were collected using BigQuery from systems including the cloud environment and training academy. He qualified these measures as proxies from owned communities rather than a count of every engaged developer.[6]

Chin traced his attention to executive alignment to an experience at an unnamed employer. His team believed its conference activities and metrics were strong, but after leadership turnover a new marketing leader responded to an event with, “well, I didn't really like the branding.” Chin said that at every subsequent company where he took a developer relations job, he interviewed with and built relationships with the CEO and leadership.[6]

Before his Neo4j work, Chin led developer relations at JFrog; earlier, he ran Oracle’s developer marketing team. In his 2025 account, he described scaling JFrog’s developer relations team through its IPO and working on DevOps, security, partner marketing, and open-source projects. These earlier roles provide context for his later work leading developer education around AI and graphs.[6][7]

His earlier technical writing includes DevOps Tools for Java Developers, coauthored with Melissa McKay, Ixchel Ruiz, and Baruch Sadogursky. Apple Books lists the O’Reilly edition as released on April 15, 2022. The collaborative guide covers the path from source code to production containers, including securing software dependencies, automating build pipelines, and maintaining software after deployment.[8]