Pharmaceutical Biotechnology M.Sc. student at Sapienza University of Rome, focused on applying AI to computational drug discovery. Currently developing an LLM-based pipeline for automated QSAR model reproduction at the Rome Center for Molecular Design (RCMD).

• 🔬 Cheminformatics: RDKit, Mordred, PaDEL, 3D-QSAR, CoMSIA, molecular docking
• 🤖 AI/ML: LLM agents, RAG pipelines, local SLM deployment, QSAR modeling
• 🖥️ Infra: managing a local GPU inference server (2× RTX 4090) - Docker, SSH, WireGuard VPN
• 🧪 Interests: computational drug discovery, fragment-based drug design, virtual screening, LLM applications in life sciences, antibiotic resistance

QSARAG LLM + RAG pipeline to automatically extract cheminformatics data from thousands of pubblications, useful for QSAR and 3DQSAR methods metanalysis or meodels reproduction. User friendly web based interface, multi-user support and chat history.

MTinyRAG Lightweight RAG system built with small language models for local, GPU-efficient inference. Cli-based interface.

Python RDKit scikit-learn Docker Linux Bash SQL Playwright BeautifulSoup UCSF Chimera AutoDock

• Portfolio / site: matteo.branchi.com
• LinkedIn: matteo-branchi-89a35b290
• Email: branchi.matteo@tutamail.com