Joel Van Horn is a New York–based Head of Engineering with 14 years of experience building data- and AI-first platforms. He currently drives technical strategy at Revic AI, architecting an agentic platform with Databricks Medallion architecture, modular LLM-enabled agents, and production-grade ML infrastructure. Previously he led NLP-to-SQL and small-model optimizations for a top-20 financial institution—achieving over 75% query accuracy—and founded engineering firms that delivered real-time analytics, event-sourced systems, and headless e-commerce platforms. Joel blends hands-on systems work with product leadership: he’s an active open-source Ruby contributor who removed ActiveSupport from a major event store, improved dotenv’s multi-line parsing, and helped shape Swagger output for Grape APIs. He’s known for translating cutting-edge LLM research into resource-efficient, auditable production systems and scaling data pipelines with Databricks and Airflow.
A Ruby implementation of an Event Store based on Active Record
Role in this project:
Back-end Developer
Contributions:182 commits, 33 PRs, 72 pushes in 1 year 7 months
Contributions summary:Joel primarily focused on refactoring the Ruby Event Store's persistence layer, specifically the integration with a ROM-based database. They removed dependencies on ActiveSupport, moved and renamed read methods, and refactored the `add_to_stream` functionality to improve efficiency. Furthermore, the user implemented JSON/YAML serialization in the adapter's schema.
A Ruby gem to load environment variables from `.env`.
Role in this project:
Back-end Developer
Contributions:16 commits, 3 PRs, 14 comments in 3 months
Contributions summary:Joel primarily contributed to improving the `dotenv` Ruby gem, focusing on parsing and handling environment variables from `.env` files. Their work involved modifying the parsing logic to correctly handle multi-line values, escape special characters, and include variables without values. Additionally, the user added tests to validate these parsing improvements and fixed Rubocop offenses. The user also addressed a regression related to how the gem parses comments and includes variables without values.
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.