My core is Ruby on Rails on the backend and React on the frontend, with Java as a solid second stack. I am moving my stack toward Elixir/Phoenix LiveView, where I have already built two products with my AI-assisted workflow. This page details each area; for where I applied them, see the projects and the experience.
Backend
- Ruby on Rails (15+ years): full-stack Rails applications, background jobs with Sidekiq, testing with RSpec, payments with Stripe, external integrations and refactoring of legacy systems.
Frontend
- React and JavaScript, for example in the Text Tools for Workdocs monday.com app.
- Hotwire (Turbo, Stimulus) for Rails, Phoenix LiveView for Elixir, and Tailwind CSS.
AI engineering
- Daily AI-assisted (agentic) software engineering with Claude Code, Codex and Windsurf, for fast system development with accountability.
- A defined workflow: I set the architecture, design principles and tools; specify each feature with expected results, examples and counterexamples; review and approve the agent’s plan; implement with checkpoints and traceability; review the code myself plus a cross-review by a second agent; and require new tests, the full test suite, linters, formatters and a security agent before staging and QA. Repeated problems become rules and repeated tasks become agent skills. See how I work.
- Integrating LLM APIs (OpenAI, Claude, Gemini, Perplexity) into products, with multi-provider fallback and bring-your-own-keys, as in Agonai.
Elixir (growing)
Elixir and Phoenix LiveView are the stack I am moving toward. I have built two products in it by directing AI coding agents with my workflow: Agonai and TechRepair, using Phoenix LiveView, Ash Framework and Oban. My hands-on Elixir level is still developing, and I am deepening it, including the fundamentals of the language and OTP.
Data & APIs
- PostgreSQL, Redis, MySQL and DB2: data modeling, query optimization and migrations, including a migration of around 7 million records from DB2 to PostgreSQL with no data loss.
- REST APIs and GraphQL (Absinthe), and microservices architectures.
- ETL with Apache Hop and dashboards with Looker Studio, as in the Fundación Telefónica KPIs.
Also strong in: Java
- Java, Spring Boot and Jakarta EE / Java EE (EJB, JPA, JSF, JMS): enterprise applications and microservices.
- Architected SIMIEC, Ecuador’s national immigration control system, serving millions of users with 99.9% uptime.
DevOps & quality
- Docker, AWS (S3, EC2, RDS), Fly.io and Heroku for deployment.
- GitHub Actions for CI/CD, Git, Linux, and automated testing with RSpec.
Architecture & practices
- Distributed systems architecture for fast, scalable, resilient, maintainable solutions.
- System integration: payment providers, third-party APIs, LLMs, HL7 healthcare protocols and specialized hardware (passport scanners, fingerprint readers).
- Legacy modernization: refactoring, performance improvements and data migrations, such as moving a billing system from Stripe to a new provider with the Adapter pattern.
- Agile delivery (Scrum, Kanban) with distributed remote teams, code reviews and mentoring.
Domains
Government systems (immigration control, transparency), healthcare (hospital information systems, HL7), competitive intelligence and AI visibility, field service operations, education KPIs, tourism ERPs, billing and payments.
Key achievements
- SIMIEC: architected Ecuador’s national immigration control system, serving millions of users with 99.9% uptime.
- Data migration: around 7 million records moved from DB2 to PostgreSQL, ETL included, with no data loss.
- Real-time notifications: introduced Turbo Streams push notifications at Dominion Global, so users are told when heavy, long-running reports are ready.
- Products built with AI-assisted engineering: Agonai and TechRepair, in Elixir, Phoenix LiveView and Ash, a stack I was new to.
Earlier technologies
Earlier in my career I also worked with C# / .NET and C++ (hardware integration modules), Oracle, Jasper Reports, Openbravo ERP and Backbone.js.