# Carmine Paolino

> I build AI tools at Chat with Work and RubyLLM. Co-founded Freshflow. Outside tech, I make music, run Floppy Disco, and take photos.

## About

Carmine Paolino is an AI engineer, open-source maintainer, and founder building AI tools at Chat with Work and RubyLLM. He previously co-founded Freshflow and also publishes music and photography.

- Role: AI engineer, open-source maintainer, and founder

## Key Topics

- Ruby
- Ruby on Rails
- Jekyll
- AI applications
- Large language models
- LLM APIs
- RubyLLM
- Open source software
- Async Ruby
- Developer tooling
- Product engineering
- Startups
- Music technology
- House music
- Photography

## Core Topics

Ruby, Rails, Jekyll, AI applications, LLM APIs, open source, async Ruby, developer tooling, product engineering, startups, music, and photography.

## Primary Projects

RubyLLM, Chat with Work, Floppy Disco, Crimson Lake, Paolino Photography, Cluster Headache Tracker, and Jekyll VitePress Theme.

## Posts

- [RubyLLM::Schema Is Now Schematist: A JSON Schema DSL for Ruby with Full Draft 2020-12 Coverage](https://paolino.me/schematist/): RubyLLM::Schema is now Schematist: a general purpose JSON Schema DSL with full Draft 2020-12 coverage, values resolved at render time, and no runtime dependencies.
- [Founding a Company in Germany: €9,600, 152 Days, and I Still Can't Send an Invoice](https://paolino.me/founding-a-company-in-germany/): I started a company in Germany in late January. By late June I had spent 9,600 euros, registered two companies, and still cannot issue a single invoice of my own. Here is the timeline, with the bill next to every step.
- [RubyLLM 1.16: Concurrent Tool Execution, Rails-Style Instrumentation, and api_base for Every Provider](https://paolino.me/rubyllm-1-16/): RubyLLM 1.16 runs your tools concurrently in threads or fibers, makes RubyLLM observable without monkey patching, and lets every native provider sit behind a proxy.
- [Engineering Is Not Dead, Because Accountability Isn't](https://paolino.me/engineering-is-not-dead/): Models can generate the code. They cannot be accountable for it. The real distinction is whether the result is owned.
- [Production Experience Cannot Be Hallucinated](https://paolino.me/production-experience-cannot-be-hallucinated/): A paid Medium article claimed hard-won production lessons about RubyLLM. The code had not even run, and the regenerated version only made the fake experience harder to spot.
- [RubyLLM 1.15: Image Editing, Cost Tracking and Less Tool Boilerplate](https://paolino.me/rubyllm-1-15/): RubyLLM 1.15 adds image editing, cost tracking, inferred tool parameters, additive callbacks, and Rails fixes.
- [kamal-backup: Scheduled Rails Backups for Kamal Apps](https://paolino.me/kamal-backup/): One Kamal accessory for encrypted Rails database and Active Storage backups, restore drills, and redacted evidence for security reviews.
- [Ruby Concurrency: What Actually Happens](https://paolino.me/ruby-concurrency-what-actually-happens/): Every 'what happens when' question about Ruby concurrency, answered with diagrams.
- [Making the Rails Default Job Queue Fiber-Based](https://paolino.me/solid-queue-doesnt-need-a-thread-per-job/): I patched Solid Queue to run jobs as fibers. Now it ships in 1.6.0, making the Rails default ready for highly concurrent, cooperative I/O-bound jobs.
- [Your Agent's Context Window Is Not a Junk Drawer](https://paolino.me/your-agents-context-window-is-not-a-junk-drawer/): Strategies to combat context rot.
- [I Built a Monitor Configuration Tool for Hyprland](https://paolino.me/hyprmoncfg-monitor-configuration-for-hyprland/): A spatial TUI editor with drag-and-drop, safe apply with revert, workspace planning, and a hotplug daemon. All in two zero-dependency Go binaries.
- [Comb Shaped Slices](https://paolino.me/comb-shaped-slices/): You don't need to pick a vertical. Build for yourself, listen to your customers, and the shape that emerges is yours.
- [Ruby Deserves Beautiful Documentation](https://paolino.me/ruby-deserves-beautiful-documentation/): The Ruby community doesn't have a great documentation theme. So I made one. Jekyll VitePress Theme brings VitePress's docs UX to Jekyll.
- [RubyLLM 1.14: From Zero to AI Chat App in Under Two Minutes](https://paolino.me/rubyllm-1-14-chat-ui/): RubyLLM 1.14 ships a Tailwind chat UI, Rails generators for agents and tools, and a simplified config DSL. Watch the full setup in 1:46.
- [Ruby Is the Best Language for Building AI Apps](https://paolino.me/ruby-is-the-best-language-for-ai-apps/): A pragmatic, code-first argument for Ruby as the best language to ship AI products in 2026.
- [RubyLLM 1.12: Agents Are Just LLMs with Tools](https://paolino.me/rubyllm-1-12-agents/): Agents aren't magic. They're LLMs that can call your code. RubyLLM 1.12 adds a clean DSL to define and reuse them.
- [Dictation Is the New Prompt (Voxtype on Omarchy)](https://paolino.me/dictation-is-the-new-prompt/): Stop typing every prompt. Speak it instead, with a fast Rust stack and a clean Omarchy setup.
- [Nano Banana with RubyLLM](https://paolino.me/nano-banana-with-rubyllm/): Nano Banana hides behind Google's chat endpoint. Here's the straight line to ship it with RubyLLM.
- [RubyLLM 1.4-1.5.1: Three Releases in Three Days](https://paolino.me/rubyllm-1.4-1.5.1/): Structured output that works, Rails generators that didn't exist, and why we shipped Wednesday, Friday, and Friday again.
- [Async Ruby is the Future of AI Apps (And It's Already Here)](https://paolino.me/async-ruby-is-the-future/): How Ruby's async ecosystem transforms resource-intensive LLM applications into efficient, scalable systems - without rewriting your codebase.
- [RubyLLM 1.3.0: Smarter Attachments, Multi-Tenancy, and No More Manual Model Tracking](https://paolino.me/rubyllm-1-3/): Attachments figure themselves out, contexts isolate configuration per tenant, and model data stays current automatically.
- [The LLM Capabilities and Pricing API is Live](https://paolino.me/standard-api-llm-capabilities-pricing-live/): The standard API for LLM model information I announced last month is now live and already integrated into RubyLLM 1.3.0.
- [A Standard API for LLM Capabilities and Pricing](https://paolino.me/standard-api-llm-capabilities-pricing/): No provider exposes model capabilities and pricing through their API. So we're building one.
- [RubyLLM 1.0](https://paolino.me/rubyllm-1-0/): One Ruby API for OpenAI, Claude, Gemini, and more. Chat, tools, streaming, Rails integration. No ceremony.
- [Building Cluster Headache Tracker from a Hospital Bed](https://paolino.me/cluster-headache-tracker/): Cluster headaches hit up to eight times a day. The hospital gave me a paper form with one line per day. So I built an app between attacks.
- [Bye Freshflow](https://paolino.me/bye-freshflow/): I co-founded Freshflow 3.5 years ago. Now I'm leaving to build something of my own.

## Links

- About: https://paolino.me/about/
- GitHub: https://github.com/crmne
- LinkedIn: https://www.linkedin.com/in/carminepaolino
- Twitter: https://twitter.com/paolino
- Instagram: https://www.instagram.com/crmne/
- Soundcloud: https://soundcloud.com/crimsonlakemusic
- Paolino: https://paolino.photography
- Crimsonlake: https://crimsonlake.live