Hello, World

I'm Matteo Carpentieri

|

Computer Engineer specializing in AI and Robotics — building systems that perceive, reason, and act at the edge of what machines can do.

View My Work Get In Touch
bash — matteo@portfolio:~
Matteo Carpentieri

About Me

I'm a Computer Engineer from Padua, Italy, with a Master's degree in Artificial Intelligence & Robotics (110/110). My work lives at the intersection of machine intelligence and real-world systems — from medical AI platforms to robotic manipulation and deep learning research.

Currently at Laife Reply, I'm building an end-to-end RAG-based healthcare platform that helps General Practitioners query clinical pathway documents through natural language — owning the full stack from document ingestion to LLM integration.

Hands-on experience across cloud, big data, robotics, and computer vision.

✉ Email in LinkedIn

Work Experience

Sep 2025 — Present
AI Consultant
Laife Reply · Padua, IT
  • Building "Punto di Accoglienza" for IOV (Istituto Oncologico Veneto) — a RAG-based chatbot letting General Practitioners query clinical pathway documents in natural language for clinical decision support.
  • Owning the system end-to-end: document ingestion, retrieval pipeline, LLM integration, and full-stack web application.
RAGLLMLangChainFull-StackHealthcare AI
Oct 2024 — Dec 2024
ML Research Intern
Profactor GmbH · Steyr, AT
  • Master's thesis research: deep learning-based defect classification of carbon-fiber composites using photometric stereo modalities.
  • Benchmarked CNN architectures combining photometric stereo with patch-overlay processing — reached 98.5% accuracy with data augmentation on a real-world industrial dataset.
  • Applied class weighting, label smoothing, and Class Activation Mapping to reduce overfitting and improve interpretability.
PyTorchCNNComputer VisionPhotometric Stereo

Education Path

Oct 2022 — Mar 2025
MSc in Computer Engineering
University of Padua · AI & Robotics

Thesis: Deep Learning-based Defect Type Classification of Carbon Fiber Composites in Photometric Stereo Feature Modalities

110/110
Oct 2019 — Nov 2022
BSc in Computer Engineering
University of Padua

Thesis: Neural Networks: General Structure and Feed-Forward Models

96/110

Selected Projects

01

Healthcare AI — Punto di Accoglienza

RAG-based chatbot for IOV letting GPs query clinical pathway documents in natural language. Built end-to-end: document ingestion, retrieval, LLM integration, web app.

RAGLangChainReactNode.js
02

Carbon Fiber Defect Classifier

CNN architectures combining photometric stereo with patch-overlay for CFRP defect classification. Achieved 98.5% accuracy with data augmentation on a real-world industrial dataset.

PyTorchCNNComputer VisionCAM
03

Autonomous Robot Navigation

ROS-based navigation for the Tiago robot. Action Client/Server architecture for autonomous movement with real-time LiDAR obstacle detection and dynamic mapping.

ROSLiDARMotion PlanningC++
04

Sport Video Analysis

Computer vision pipeline to identify and track players across teams, generate semantic game information, and understand field boundaries and action context.

OpenCVObject TrackingSegmentationPython
05

ML vs GCNN — Accident Severity

Comparative analysis of ML models and Graph Convolutional Networks for accident severity classification using road intersection data from London.

GCNNScikit-learnGraph MLPython
06

Robotic Pick & Place — SAGRONE

Pick-and-place system using ROS and MoveIt with AprilTag detection and motion planning for autonomous object fetching and delivery in a simulated environment.

ROSMoveItAprilTagPython

Technical Skills

AI & Machine Learning
PyTorchPyTorch
TensorFlowTensorFlow
Scikit-learnScikit-learn
OpenCVOpenCV
LangChainLangChain
HuggingFaceHuggingFace
Machine Learning Deep Learning Computer Vision Generative AI RAG Systems
Programming Languages
PythonPython
C++C++
JavaJava
JavaScriptJavaScript
PHPPHP
HTML5HTML5
CSS3CSS3
SQLSQL
Web & Cloud
ReactReact
Node.jsNode.js
GCPGoogle Cloud
SparkApache Spark
DockerDocker
GitGit
MapReduce Data Streaming Database Design
Robotics
ROSROS
Autonomous Navigation LiDAR Processing Motion Planning MoveIt 3D Data

Let's Connect

Open to opportunities, research collaborations, and interesting conversations.

✉ Email Me in LinkedIn