Available for new opportunities
Software Engineer & Tech Lead · Ruby on Rails · Rust · Elixir/Phoenix · AI Engineering · MCP

Ismael
Marin

Software Engineer & Tech Lead · León, Guanajuato, Mexico 🇲🇽

Software engineer and technical leader with deep experience building backend systems, SaaS products, integrations, and engineering teams for U.S. and Latin American companies. Ruby on Rails is my deepest production ecosystem, while I now work regularly with Rust, Elixir/Phoenix, and Python, particularly around developer infrastructure and AI systems. I approach AI as a software engineering problem: clear specifications, useful context, tests, guardrails, and evals. I have worked remotely throughout my career and remain hands-on with architecture and implementation.

Portfolio, résumé & contact Code & projects
ismael@portfolio:~
ismael@portfolio:~ $ about
Ismael Marin · Software Engineer & Tech Lead León, Guanajuato, Mexico Ruby on Rails · Rust · Elixir/Phoenix · Python · AI Engineering Open to full-time, part-time, and contract opportunities. Type 'help' to explore.
ismael@portfolio:~ $
Quick commands:
scroll
370+
Merged pull requests across 20+ production repositories at Dealerware · May 2022–April 2026
40+
ADRs/RFCs authored at Dealerware · May 2022–April 2026
~10,000
Telematics events per minute at Dealerware · May 2022–April 2026
Open Source

Featured Projects

ruby-skill-bench

Learn more →

An evaluation framework for measuring whether engineering skills and additional context improve AI-agent performance, using isolated runs, baseline-versus-context comparison, and blind judging.

RubyGems More than 3,000 total downloads
  • Side-by-Side Evaluation: Quantify the impact of specific context or skills on agent performance
  • Isolated Git Sandboxes: Compare runs with clean diffs and repeatable setup
  • Multi-Provider: Native support for Anthropic, Gemini, OpenAI, DeepSeek, Groq, Ollama, and more
  • LLM-based blind judging across Correctness, Skill Adherence, Code Quality, and Test Coverage
$ gem install ruby-skill-bench
Ruby AI Benchmarking Evaluation Git Sandboxing

rails-ai-bridge

Learn more →

MCP and context infrastructure that lets AI coding assistants understand real Rails applications through generated project context and read-only introspection.

RubyGems More than 10,000 total downloads
  • Generates static context files for every AI tool — CLAUDE.md, AGENTS.md, .cursorrules, copilot-instructions.md, .windsurfrules, .mcp.json
  • No server required — commit the files, whole team benefits automatically
  • Live MCP server mode for real-time Rails introspection on demand
  • Works with Claude Code, Cursor, Codex, GitHub Copilot & Windsurf
$ gem 'rails-ai-bridge'
$ rails generate rails_ai_bridge:install
$ rails ai:bridge
Ruby MCP Claude Code Cursor Codex Copilot

agent-mcp-runtime

Safe Rust CLI
Learn more →

Safe Rust CLI for Model Context Protocol (MCP) runtime management. Orchestrates communication between LLMs and external tools via standard I/O streams using the ReAct reasoning pattern.

  • Memory safety with strict compile-time checks and zero unsafe blocks
  • Async operations via Tokio for high-performance concurrent tool execution
  • Multi-provider support: OpenAI, Claude, Gemini, and Groq out of the box
  • Cross-platform compiled binary with easy-install script
$ curl -sSL https://github.com/igmarin/agent-mcp-runtime/releases/latest/download/install.sh | sh
Rust MCP CLI Systems Programming

pi-vida

Agent Launcher

A launcher for Pi and other coding-agent hosts that loads language-specific skills, safety controls, and workflows for Ruby, Rust, Python, and Elixir.

  • Choose a language profile and host from the target repository
  • Use solo, plan-build-review, or specialist-team workflows
  • Apply safety gates and checks before changes are made
$ pi-vida ruby
Pi Agent Workflows Ruby Rust

brigid

Rust CLI

A Rust CLI that turns a codebase into a multi-chapter Markdown tutorial with Mermaid diagrams, helping people understand an unfamiliar system.

  • Maps abstractions and their relationships from source code
  • Generates setup, architecture, and chapter guides with real file paths
  • Checkpoints long runs and can resume or update changed sections
$ cargo install brigid-cli
Rust LLM Tooling Developer Onboarding

rs-guard

AI Code Review CLI

A Rust CLI for multi-provider AI reviews of GitHub pull requests, with a local pre-commit mode for reviewing changes before they are committed.

  • Runs in GitHub Actions or locally with configurable project rules
  • Supports multiple hosted and local LLM providers
  • Exports per-run token, latency, and cost-estimate metrics
$ cargo install rs-guard --locked
Rust GitHub Actions Code Review Pre-commit
Skills · Ruby · Rust · Elixir/Phoenix · Hanami · AI planning

rails-agent-skills

Learn more →

Curated AI agent skill library for Ruby on Rails development. Structured SKILL.md files that give coding agents the context and conventions to generate production-quality Rails code.

npx skills add igmarin/rails-agent-skills
Ruby on Rails AI Agents TDD Cursor Claude Code

hanakai-yaku

Learn more →

Curated library of atomic development skills and orchestrated agents for AI coding assistants, enforcing TDD and DDD discipline for Hanami 2.x Ruby applications.

npx skills add igmarin/hanakai-yaku
Hanami Ruby DDD Clean Architecture

ruby-core-skills

Learn more →

Core Ruby skills for AI agent evaluation. A structured library of language fundamentals, standard library mastery, and idiomatic patterns for benchmarking.

npx skills add igmarin/ruby-core-skills
Ruby AI Evaluation Language Skills

rust-core-skills

Skills for idiomatic Rust development, covering ownership, type-driven design, error handling, project context, and Cargo TDD.

npx skills add igmarin/rust-core-skills
Rust Ownership Cargo TDD

elixir-phoenix-skills

Skills and workflows for idiomatic Elixir and production Phoenix work, including LiveView, Ecto, OTP, testing, security, and deployment.

npx skills add igmarin/elixir-phoenix-skills
Elixir Phoenix LiveView Ecto

agnostic-planning-skills

Learn more →

Framework-agnostic AI planning capabilities. Enforces PM and software engineering planning discipline with structured templates for PRDs and task estimation.

npx skills add igmarin/agnostic-planning-skills
AI Agents Planning Cross-platform Agentic Workflows
Capabilities

Tech Stack

Core Infrastructure
Ruby on Rails Rust Elixir/Phoenix Python PostgreSQL Redis Cloudflare Vercel AWS Docker
AI & Tooling
AI Agents Context Engineering Evals Guardrails RAG Spec Driven Development LangGraph Local Models LLM Integration MCP Multi-provider LLM Systems
Architecture
API Design Domain-Driven Design Distributed Systems Spec-Driven Design Monolith Design Hexagonal Architecture
Collaboration

How we can work together

🏗️

AI Integration & Tooling

I help teams safely integrate LLMs into existing Rails monoliths, build MCP servers, and setup agent-based workflows.
👨‍💻

Staff Augmentation

Available for Senior/Lead level feature development, capable of driving complex technical initiatives from PRD to production.
📐

Architecture Consulting

System design, performance optimization, and ADR authoring for scaling Ruby on Rails applications.
Writing

Latest articles

🤖

How I Built a Composable AI Skill Ecosystem for Ruby with Help from 4 Different AI Agents

Published · medium.com/@igmarin
→
✍️

From Static Prompts to Production Agents: My Journey with Tessl and AI Skills

Published · medium.com/@igmarin
→
✍️

Your JSON Column Is Not a Settings System

Published · medium.com/@igmarin
→
✍️

The Engineer Who Understands the Product Will Always Win

Published · medium.com/@igmarin
→
🤖

Beyond Vibe Coding: How rails-ai-bridge Achieved "Semantic Parity"

Published · medium.com/@igmarin
→
🤖

Stop "Vibe Coding": How I Bridged the Gap Between Rails and AI

Published · medium.com/@igmarin
→
All articles on Medium →
Media & Speaking

Thought Leadership

De Mentes Aumentadas Podcast Logo
Bi-weekly AI Radio Show Spanish 🇪🇸

De Mentes Aumentadas

A syndicated bi-weekly segment about Artificial Intelligence and technology broadcasted on national FM radio, syndicated online as a podcast. We break down the implications of agentic workflows, machine learning models, and complex system architectures to bridge technical depth with public understanding.

Follow on Spotify
Featured Episode

¿Nos escucha el celular? / Is our cellphone listening to us?

In this segment, we analyze and demystify the common belief that our phones actively wiretap our daily audio. I explain how data brokers, web tracking networks, cookie syncing, and predictive machine learning models analyze non-audio signals to predict user desires with hyper-accuracy, mimicking active listening.

Listen to the Episode →
Connect

Get in touch

Open to full-time, part-time, and contract opportunities in software engineering and AI engineering.

Download Resume (PDF) Let's Talk
🌐
Website
ismaelmarin.dev
💼
LinkedIn
ismaelmarin
⌨️
GitHub
igmarin
📝
Medium
@igmarin
✉️
Email
ismael.marin@gmail.com
🧠
Rails Skills
AI Agent Skills & Workflows
💎
RubyGems
rails-ai-bridge
📊
Skill Bench
AI Evaluation Engine