Salvatore Iavarone

Focused on large language model systems and agentic architectures, with emphasis on tool-using LLMs, retrieval-augmented generation, and multi-step reasoning pipelines. Actively working on LLM agents for the Minerva models from Sapienza NLP, exploring robust orchestration of workflows (planning, execution, memory) and evaluation of generated outputs. Strong interest in bridging classical software engineering with modern neural systems, especially in production-grade AI systems, automation, and embedded intelligence in real-world applications.