Open to opportunities

Sergio Medina Muñoz

@sergiomedina

Message

I build reliable local RAG systems with traceable, accurate answers and verifiable citations.

Spain

Message

What I'm looking for

I'm looking for AI/ML Engineer opportunities, particularly in RAG, LLMs, NLP, and applied ML. I value building AI that is accurate, efficient, private, and reliable, and I'm open to remote roles across Spain, Europe, and worldwide.

I independently built and evaluated GRAIA, a scalable fully local RAG system for academic use cases, as my Computer Engineering thesis at the University of Granada. It earned a 10/10 with the committee's distinction.

I designed its web and PDF ingestion, hybrid dense-embedding and BM25 retrieval, Reciprocal Rank Fusion, cross-encoder reranking, MMR diversification, and local Llama 3.1 8B generation through Ollama. I also built out-of-domain and faithfulness safeguards plus citation verification, then benchmarked the system with 74 questions; it improved factual accuracy from approximately 5% to 87%, delivered 100% verifiable citations, and maintained sub-five-second latency on a consumer GPU.

Experience

Work history, roles, and key accomplishments

AI/ML Engineer

GRAIA

Independently designed, built, and evaluated GRAIA, a scalable, fully local RAG system for the academic domain, achieving 10/10 with distinction. Focused on reliability with out-of-domain gates, faithfulness safeguards, and citation verification.