This month we teamed up with our friends at GDG Phoenix / PHX Android for a hands-on build night at HeatSync Labs.
Rather than a passive lecture, this was an open working workshop and hack night focused on what developers can build with AI without spending a dime – running models locally, exploring generous free tiers, and sharing practical tooling across the Phoenix tech community.
Talks
Ruby AI Gems: Building AI Applications in Ruby — Marc Heiligers
Marc gave an overview of the evolving AI ecosystem in Ruby, showing how developers can build production-ready AI features, RAG pipelines, and context-aware agents natively in Ruby without vendor lock-in.
- RubyLLM: A unified interface across 17+ LLM providers (OpenAI, Anthropic, Gemini, etc.) offering clean Ruby ergonomics, streaming responses, and structured JSON outputs out of the box.
- Ruby MCP Client: Standardizing context integration via the Model Context Protocol, connecting AI models directly and securely to local files, tools, and databases.
- Andrew Kane’s ML Gems: Bringing native machine learning directly to Ruby with Torch.rb (deep learning & neural networks), Informers (native transformer models for embeddings and sentiment analysis), and Pgvector-ruby (vector similarity search for RAG in Active Record).
Presentation — Sheldon McGee
Sheldon McGee co-presented for the evening, sharing practical insights and demos around building with free developer AI tooling.
Links
- RubyLLM — Unified interface for AI APIs in Ruby and Rails
- Model Context Protocol: Build a Client in Ruby
- 16 New ML Gems for Ruby by Andrew Kane
- More ML Gems for Ruby by Andrew Kane
Thanks
- Co-hosted with GDG Phoenix / PHX Android
- Event space and hosting provided by HeatSync Labs