I build production-ready AI architectures and data pipelines focused on ROI and operational efficiency. Currently specialized in Agentic AI, MLOps, and FinOps.
- AI/LLMs: OpenAI, Claude, DeepSeek, LangChain, CrewAI, LangGraph.
- Data & Tooling: Python (Async/FastAPI), DuckDB, Dremio, Airflow, SQL.
- Architecture: RAG, Model Context Protocol (MCP), Vector Databases, Txt2SQL.
- Cloud & DevOps: Oracle Cloud (OCI), Docker, GitHub Actions.
- Review CLI: A developer tool (DX) that automates code reviews using AI agents. Built with Python (Typer/Rich) and optimized for low-friction context injection.
- Txt2SQL Engine: Scalable platform for natural language database querying with process isolation via DuckDB.
- MCP Proxies: Implementing the Model Context Protocol to bridge legacy systems (Delphi/SQL) with modern LLMs.
- Nota Paraná Automation: High-performance web scraping using Playwright with asynchronous concurrency for data extraction.
- 🔭 Currently working on: AI Agentic Workflows for Enterprise SaaS.
- 🎙️ Speaker/Mentor: Translating AI complexity into business value.
- 💬 Ask me about: LLM integration, FinOps, and why automation must be pragmatic.

