Georgios Peikos

I am an Applied AI Engineer and Search/NLP Researcher at the University of Milano-Bicocca, with a background in Electrical and Computer Engineering. My work focuses on Large Language Models, Retrieval Augmented Generation, and Information Retrieval, with applications in healthcare and other high reliability domains.

I design and deploy end-to-end AI systems that combine data pipelines, semantic retrieval, reranking, model training, inference, and evaluation. My recent work includes evidence-grounded RAG systems, conversational search, clinical trial retrieval, and trustworthy AI workflows.

I have experience training and adapting small and mid-scale language models, including 1B-7B parameter models, using PyTorch DDP on Slurm-managed HPC infrastructure. System design is guided by real-world constraints such as evidence grounding, privacy, efficiency, and expert oversight.

I also teach Information Retrieval and Recommender Systems at Bachelor’s and Master’s level, with an emphasis on evaluation and practical system building.

I am open to collaborations and technical roles involving the design and evaluation of applied AI systems, particularly in healthcare, enterprise search, legal search, and other domains where reliability, grounding, and evaluation matter.