The September SF Ruby meetup was hosted at Binti, and it was great to see a new company join the roster of Ruby meetup hosts.
The evening kicked off with Chris Fung introducing Binti and their mission. Binti builds a platform for child care agencies with the goal of making sure every child gets a home. Recently, they added AI features to help social workers work more efficiently—so much so that they were even mentioned by Anthropic's CEO in Forbes.
Next, Irina spoke about the upcoming SF Ruby Conference. I'm excited to attend The Ruby Passport will be available there. Tickets currently have a $50 discount,
While talking about the recent Ruby Central changes, I especially liked her quote: "We can choose to believe that every person is actually acting from the best possible intentions." It was a refreshing reminder of the positive spirit in the Ruby community.
The talks then continued:
- André Arko introduced rv.dev, a new Ruby version manager inspired by uv and built in Rust.
- Cade Friedenbach gave a talk on how Binti built their webhooks and data export platform. They kept things simple with Service Objects, Wisper, and background jobs, and used JSON schema validation before saving to the database via activerecord_json_validator.
- Kody Kendall spoke about llamapress, an open-source platform for building Ruby on Rails apps with LLMs. He explained how Rails runner can be used as a tool for agentic coding, and why Rails conventions and concise syntax make it especially suited for AI-driven development.
- Arjun shared how Ruby is being used in medical software to build 3D models from 2D scans. It was exciting to see Ruby applied outside of web development. He also talked about the importance of compliance and testing in the medical field—no "move fast and break things" there.
- Finally, Pranav from Chatwoot presented their open-source gem for building AI agents based on Ruby LLM. This is the foundation behind Captain, their agentic customer support assistant. It enables customers to create their own assistants by providing custom instructions to triage support tickets. I liked how simple their API was. After the talk, I got to chat with him about Chatwoot's journey and how they've kept their architecture and deployment simple.
Overall, it was a fantastic meetup. It was great to see the Ruby community experimenting with AI in practical ways, beyond the hype.