I build AI-powered backend systems and LLM applications, with experience spanning RAG pipelines, multi-agent workflows, NLP, and cloud infrastructure.
My work focuses on turning research-driven AI ideas into reliable products, including fraud detection systems, enterprise messaging platforms, and Text-to-SQL agents. I’ve worked across both startup and research environments, contributing to deployed systems as well as published NLP research accepted at SemEval 2026 (ACL Workshop).
- 💼 Current: AI Engineer @ EstateBotics & Co-Founder @ Remy.AI
- 🎓 Education: Data Science at University of São Paulo (USP) & Harvard Computer Science in Brazil at Fundação Estudar.
- 🛡️ Focus: Architecting scalable RAG pipelines, NLP agents, and production-grade ML systems.
- Remy.AI – Real-time fraud detection chatbot for WhatsApp & Telegram using NLP and LLMs.
- TextToInsight - Text-To-SQL library with a next step of insight, include GraphRAG pipelines to filter Schema databases and rn aim to Host local models to achieve big results at low costs. Reduced the tokens usage in ~25% using GraphRAG.
- Unfaker – Chrome extension for fake news detection using Transformers.
- ML-TutorBot – Multilingual RAG tutor with code interpretation; reduced learning time by ~20%.
- ML Model Factory – Automated MLOps pipeline using MLflow & Docker; reduced Data Scientist setup time by 33%.
- Core: Python, C++, R, Java, C, Javascript
- AI, NLP & ML: LangGraph, LangChain, Transformers, PyTorch, RAG, OpenAI/Gemini APIs, TensorFlow, Scikit-learn, PEFT, TRL, vLLM, Unsloth, HuggingFace
- Ops & Cloud: Docker, MLflow, AWS, Azure, FastAPI
- Databases: SQL (PostgreSQL, MySQL), NoSQL (DynamoDB, SQLite)
