MeteoRAG
Intelligent weather assistant with RAG combining real-time INMET API data with LLMs for natural language queries.
The Problem
Weather data scattered across technical APIs without user-friendly interfaces, making quick queries about current and historical conditions difficult.
The Solution
Complete ingestion pipeline with deterministic chunking, semantic retrieval and LLM response generation. Deployed via Kubernetes with CI/CD on GitHub Actions, monitoring with Prometheus and Grafana.
Result
End-to-end RAG pipeline in production with automated Kubernetes deployment, GitHub Actions CI/CD, Prometheus/Grafana monitoring and natural language weather responses for MG.
Note
This is the MVP for a much larger project: a large-scale RAG-powered AI platform with DuckDB that will generate insights on rainfall, landslides and alerts across the entire Zona da Mata region in Minas Gerais — cross-referencing data since 1980 with predictive risk analysis. The repository is open: feel free to contribute or use this RAG as a foundation for your own project.
Related Projects
JOX — ETL with LLM
Intelligent ETL pipeline for structured data extraction from agricultural bulletins in PDF using LLM with structured output.
Melisso AI Agent
AI conversational assistant embedded in the portfolio — Claude Haiku streaming, Redis rate limiting and animated interface.
CRM Analytics — Multi-tenant AI
Multi-tenant commercial intelligence platform — ML-driven customer scoring, GPT-4 insights and real-time recurrence analysis.