Tiago Antunes // AI Engineer
AI engineer building production LLM systems on Rails and Python — RAG, function-calling agents, MCP servers, and multi-model pipelines. 4+ years shipping production software, the last two focused on applied AI. Currently running four AI SaaS products (Tablrr, Tripr, ApplyTrack, SkillPro for Universities).
ls -l featured/
tablrr ★
Dependencies: Rails 8, Hotwire, Turbo Streams, Tailwind 4, PostgreSQL, Solid Queue, Solid Cache, Redis, Stripe, MCP, RubyLLM, Kamal
Rails 8 SaaS for restaurant reservations and menu management. Multi-tenant, with a slot-based availability engine, live floor management over Turbo Streams, AI menu extraction, and two MCP servers written from scratch so external AI agents can run the restaurant as a tool.
tripr ★
Dependencies: Rails 8, Hotwire, Tailwind, PostgreSQL, RubyLLM (Gemini 3.1, DeepSeek V4, GPT-5 Mini), Apify, Solid Queue, Cloudflare R2, Kamal
AI-powered travel itinerary builder. A multi-model LLM pipeline generates, fact-checks, and ranks attractions, then a constraint-based scheduler packs them into day timelines respecting meal windows, travel time, and capacity budgets.
applytrack ★
Dependencies: Rails 8, Hotwire, Turbo Streams, Tailwind, PostgreSQL, Devise, RubyLLM, Chrome Extension (Manifest V3), Adzuna, JSearch, Remotive, Stripe, Kamal
Rails 8 job application tracker shipped as a web app plus a Chrome extension, both served by one backend. Its AI assistant doesn't just answer — it calls tools to build applications from a LinkedIn URL, score how well a CV matches a job, and write the cover letter.
skillpro for universities ★
Dependencies: Rails 8, Hotwire, Turbo Streams, PostgreSQL, OpenAI, Structured Outputs, Solid Queue, Active Storage, Kamal
White-labelled, multi-tenant platform where universities run AI skill assessments across their courses. An AI examiner holds a real conversation with each student, grades them against a taxonomy built from the professor's own course materials, and hands the instructor per-subskill analytics and CSV exports.
cat about/README.md
Hi, I'm Tiago Antunes — an AI engineer specialising in production LLM systems: RAG, function-calling agents, multi-model orchestration, and structured outputs. 4+ years of production engineering in Ruby on Rails and Python, the last two focused on applied AI. M.Sc. Software Engineering, University of Coimbra.
Spent three years at HYP shipping AI products from zero to production: built TopCashback's RAG-based content classifier with pre-filtering that kept inference costs flat as volume grew, and SprintCV's AI suite (CV import with structured outputs across multiple LLM providers, candidate-to-job fit classification, JD-tailored CV generator, conversational assistant with function-calling tools). Then at Koinly, building blockchain ingestion pipelines across messy third-party APIs.
Currently running four AI SaaS products end-to-end: Tablrr (restaurant platform with two MCP servers written from scratch, so external AI agents can run the restaurant), Tripr (travel planner routing each pipeline step to a different model, with a scheduler that fits a day around meals and travel), ApplyTrack (job tracker as web app + Chrome extension, with an assistant that acts through tools), and SkillPro for Universities (white-labelled AI examiner universities run across their courses).