Leonardo Peroni | Applied AI · Reinforcement Learning

About Me

I am a Senior AI/ML Engineer with a Ph.D. in Telematic Engineering from Universidad Carlos III de Madrid. I design and deploy end-to-end AI systems, operating across model-centric machine learning work and system-level AI engineering, from experimentation to production.

My work spans classical machine learning and deep learning, deep reinforcement learning, and LLM-based systems. I am particularly interested in hybrid roles where hands-on modeling and experimental prototyping are tightly coupled with robust productionization, including MLOps practices and reliable deployment pipelines.

Earlier in my career, I worked in applied AI research for adaptive video streaming, focusing on Quality of Experience (QoE) modeling, reinforcement learning–based adaptive bitrate (ABR) algorithms, video encoding, and transport-layer dynamics.

Experience

Senior AI/ML Engineer, Indra Sistemas – Cyberdefence
Madrid, Spain | May 2025 – Present
Research Assistant, IMDEA Networks Institute
Madrid, Spain | Mar. 2020 – Aug. 2024
Technology Consultant, Hesplora s.r.l.
Florence, Italy | Sep. 2018 – Feb. 2020

Education

Ph.D. in Telematic Engineering, Universidad Carlos III de Madrid
M.Sc. in Mechatronic Engineering, Polytechnic University of Turin
B.Sc. in Informatics and Automatic Engineering, Sapienza University of Rome

Focus Areas

Deep reinforcement learning and control with neural architectures, including value-based, policy-gradient, and actor–critic methods, as well as multi-agent systems
End-to-end AI system development, spanning model design, training and evaluation, deployment, and operational monitoring
LLM-based systems, with focus on prompting and reasoning strategies, fine-tuning, tokenization, and retrieval-augmented generation
Hybrid AI systems that combine learning-based decision-making with explicit reasoning and action execution

Current Interests

I am currently focused on extending my deep reinforcement learning background toward LLM-based and agent-oriented AI systems. I am exploring how large language models, advanced prompting and reasoning techniques, and retrieval-augmented generation can support decision-making, planning, and adaptive behavior in complex environments.