Senior Full-Stack Engineer with 8+ years building production software in Ruby on Rails, Python, and React, with the last three years focused on shipping AI and LLM-powered systems to production.
- 🛤️ Building and scaling Rails applications since version 3 — performance, reliability, and code that ages well
- 🔬 Shipping production RAG systems with LangChain, Claude, and vector databases
- ☁️ Deploying solutions on AWS, where I'm certified (Cloud Practitioner + AI Practitioner)
- 📊 Recently completed a Master's in Data Science at CEUPE — classical ML, statistical modeling, and applied analytics
- ✍️ Writing about when to use AI and when not to
Backend: Ruby on Rails (v3–8) · Python (FastAPI) · Node.js (NestJS) · PostgreSQL · Redis · Sidekiq · GraphQL
AI/ML: LangChain · Claude · OpenAI · ChromaDB · RAG pipelines · XGBoost · evaluation harnesses
Frontend: React · TypeScript · Next.js · Ember.js · Hotwire/Turbo
Cloud: AWS (Lambda, Step Functions, S3, SQS, SNS, ECS, RDS, Glue) · Docker · CI/CD · GitHub Actions
- 🟧 AWS Certified AI Practitioner
- 🟧 AWS Certified Cloud Practitioner
- 🎓 Master's in Data Science — CEUPE (completed)
- 🎓 ML Specialization — Stanford / DeepLearning.AI (completed)
🔍 AWS Docs RAG Assistant — Self-hosted RAG with hybrid retrieval (BM25 + dense embeddings), evaluation harness, semantic caching. FastAPI, ChromaDB, Redis, AWS.
🤖 Autonomous Research Agent — Anthropic tool-use protocol with guardrails: token budgets, iteration caps, loop detection. Reduced runaway runs from ~116K to ~11.5K tokens.
🏗️ LandingHub — Rails monolith connected to Claude via MCP for generating landing pages from client specs. Stripe subscriptions with idempotent webhook processing.
- ✍️ Medium
- 📧 henrylofiego@gmail.com
- 🌎 Based in Venezuela · Open to remote roles globally
"Match the tool to the problem, not to the trend. That's where I do my best work."