Skip to content
View rricc22's full-sized avatar

Highlights

  • Pro

Block or report rricc22

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
rricc22/README.md

👋 Hi, I'm Riccardo Castellano

MSc AI @ CentraleSupélec | Materials Engineer | TensorFlow Certified Developer


💫 About Me

From the foundry floor to neural networks - I build AI that solves real manufacturing problems.

Spent 3 years at Howmet Aerospace developing gas turbine components, where I saw firsthand how quality issues impact production. Now pursuing MSc AI at CentraleSupélec to bring cutting-edge ML to aerospace manufacturing.

Specialized in materials characterization through deep learning: built YOLO-based systems for steel microstructure classification, CNNs for casting defect detection, and Physics-Informed Neural Networks that solve computational physics problems 3.21× faster.

🔬 Upcoming: Internship at SAFRAN TECH POLE focusing on microstructure prediction

🌍 Franco-Italian | Fluent in 🇮🇹🇫🇷🇬🇧 (TOEIC 910, IELTS 7/9) + 🇪🇸🇯🇵

📫 Looking for: AI/ML opportunities in aerospace, manufacturing, and materials science


🌐 Socials

LinkedIn Portfolio Email


💻 Tech Stack

AI/ML Frameworks

Python TensorFlow PyTorch Keras scikit-learn

MLOps & Cloud

Google Cloud Kubernetes Docker

Data Science & Computer Vision

OpenCV Pandas NumPy Plotly

Web Development & Tools

Flask HTML5 CSS3 Jupyter Git


🚀 Featured Projects

Developed conditional PINNs solving 2D heat diffusion with 3.21× performance improvement over image-based approaches. Systematic comparison of CNN vs MLP vs PINN architectures.

YOLO-based system for automated Martensite/Pearlite identification in steel samples. Full-stack deployment with Flask.

LSTM & Transfer Learning achieving MAE: 8.94 BPM on physiological time-series data.

Strategic game AI using Alpha-Beta pruning with smart move filtering. Production ready!

Multi-objective optimization for sales territory assignment using NSGA-II and Simulated Annealing.

Formula 1 data analysis and visualization using Python, exploring race strategies and performance metrics.


🎓 Certifications

  • 🥇 TensorFlow Developer Certificate - Google
  • 📚 Google ML Engineer Path (8/12 badges completed)
  • 🎓 Deep Learning Specialization - DeepLearning.AI (Coursera)

📊 GitHub Stats




🎯 Current Focus

  • 🔬 Preparing for internship at SAFRAN TECH POLE (microstructure prediction)
  • 📚 Advancing in Google ML Engineer Path
  • 🤖 Exploring robotics control and physics-guided machine learning
  • 📊 Building production-grade ML systems with MLOps best practices

💡 Fun Facts

  • 🏃‍♂️ Marathon runner - trained ML models to predict heart rate during runs
  • 🏎️ F1 enthusiast - analyzing race data with Python
  • 🗾 Studied Japanese with 3-month immersion (100+ hours)
  • 🎮 Built strategic game AI that beats human players

📈 Profile Views


Pinned Loading

  1. SUB_3H_42KM_DL SUB_3H_42KM_DL Public

    Heart Rate Prediction from Activity Data using LSTM & Transfer Learning | Deep Learning Course Project @ CentraleSupélec | MAE: 8.94 BPM

    Jupyter Notebook 1

  2. heat-equation-ml heat-equation-ml Public

    Jupyter Notebook