Self-Corrective RAG (Self-CRAG) with LangGraph for multi-turn dialogues. Llama 3.1 8B with 4-bit quantization, hybrid retrieval with Parent-Child chunking and Cross-Encoder reranking.
Hi, I'm Marco.
AI Engineer focused on Generative AI and Large Language Models. M.Sc. in AI & Data Analytics at Politecnico di Torino. Currently researching Foundation Models for automated Data Engineering at DATA Reply.
Intern
intelligent together.
About me
AI Engineer · Politecnico di Torino · Turin, Italy
I'm a Computer Engineer specializing in Generative AI and Large Language Models, with a strong foundation in AI, Machine Learning, and Data Analytics. Currently pursuing my Master's degree in AI & Data Analytics at Politecnico di Torino.
I specialize in fine-tuning and quantizing LLMs, building RAG pipelines, and developing agentic AI systems using LangChain and LangGraph. Passionate about making Foundation Models practical and accessible.
Journal: SUAI Bulletin of the UNESCO Chair, pp. 154–163
Date: April 2024
Work Experience
Professional history in AI research and engineering
Researching Foundation Models and Serverless architectures to automate Metadata Management and Semantic Discovery.
Recruitment and member engagement for the electrical & computer engineering honor society. Organized technical events and workshops.
Conducted research on dynamic gesture recognition using RNNs for HCI. Built real-time recognition systems on Leap Motion sensor data.
Featured Projects
AI, Machine Learning, and Human-Computer Interaction
Controllable abstractive summarization via semantic supervision. Multilingual pipeline (IT/EN) with Llama and Qwen. LoRA adapters on HuggingFace.
Dynamic hand gesture recognition — BiLSTM with Attention + encoder-only Transformers. 400 recordings, 4 gestures.
Bachelor's thesis: Bidirectional RNN for complex dynamic gesture classification from time-series sensor data in real-time, with applications in healthcare and entertainment.
Real-time face detection bot with user-friendly Flet GUI for remote detection and Telegram notification. Built with OpenCV for efficient processing.
Academic Background
Education, coursework, and certifications
| Course | CFU | Grade |
|---|---|---|
| Data Science & DB Technology | 8 | 30L |
| Computer Architectures | 10 | 30 |
| Web Applications I | 6 | 30 |
| System & Device Programming | 10 | 30 |
| Software Engineering | 8 | 29 |
| Big Data Processing | 6 | 28 |
| Robot Learning | 6 | 28 |
| Large Language Models | 6 | 27 |
| ML & Pattern Recognition | 6 | 26 |
| Deep NLP | 8 | 25 |
| Explainable AI | 6 | In progress |