Within the Agorai Innovation Hub we are looking for a talented and proactive resource, who will work directly with the AI Architect and Head of Technology on use case definition and technical direction.
The researcher will work at the intersection of language understanding, knowledge representation, and trustworthy AI. This means combining LLMs with structured knowledge to ground reasoning; and building information extraction systems whose outputs can be audited.
Own the research-to-production arc on language understanding and knowledge-grounded AI: from literature review and experimentation to deployment and monitoring
Design AI systems whose outputs can be explained, audited, and trusted in regulated contexts — and know when simpler approaches serve that goal better than complex ones
Build production-quality services around your models: Design evaluation frameworks that go beyond benchmark metrics — testing for failure modes, edge cases, and behavior under distribution shift
Work on information extraction, semantic understanding,
and structured knowledge problems grounded in real industrial data
Stay current with the research frontier in language models, interpretability, and trustworthy AI
Work on a genuinely hard and open research problem with direct industrial application
The role is based in Trieste and includes a hybrid work model, offering flexibility through remote working days.
Requirements Master's degree or PhD in Computer Science, Artificial Intelligence, Data Science, Engineering or a related STEM field
At least 3–5 years of experience working with NLP or language understanding systems in production environments, such as text classification, information extraction, semantic search, RAG, or generative systems
At least 3–5 years of experience in applied NLP or computer vision, working with real-world data and operational constraints rather than benchmark-only environments
Experience working across one or more of the following sectors: LLM-based systems,
📌 Research And Innovation Engineer (Italia)
🏢 Altro
📍 Italia