New account AI focused.

About Me 🙋

• 🧑‍💻 AI engineer | ex Blockchain (@ freelance) and Backend (@ HiWayMedia) engineer
• 🚀 currently learning/experimenting/working with ML, Deep Learning and more
• 🎓 MSc AI Engineering | end September 2026 @ ProfessionAI - BA Philosophy, tech oriented, @ Milan University
• 📧 Reach me at: alceblockdev@gmail.com
• 🏡 Valencia (🇪🇸) Sydney, AUS (🇦🇺) Milan (🇮🇹)

For more:

What you can find here:

• AI Engineering Projects. A collection of applied projects covering Machine Learning, Deep Learning, Computer Vision, Generative AI, XAI, Reinforcement Learning and MLOps.

Projects:

• XAI-banking-compliance - Explainable AI (XAI) - XAI auditing framework evaluating a fine-tuned DenseNet-121 architecture on MNIST via multi-modal saliency maps, paired with an inherently transparent quadrant-density Decision Tree.
• power-plant-security-augmentation - Generative AI - Advanced pipeline leveraging multi-modal text-to-image synthesis to automatically expand training datasets and structurally improve deep vision classifier robustness under strict data-scarcity constraints.
• reinforcement-learning-cybersecurity - Reinforcement Learning - Advanced network defense simulation engine applying SARSA and Double Deep Q-Network (DDQN) agents within the gym-idsgame environment to automate threat mitigation and safeguard critical healthcare infrastructure.
• flower-recognition-agritech - Computer Vision - End-to-end computer vision classifier deploying a pre-trained ConvNeXt Tiny architecture with customized multi-layer vocabulary matching to automate fine-grained botanical categorization for smart farming.
• food-image-classification - Deep Learning with PyTorch - Deep transfer learning pipeline in PyTorch optimizing pre-trained VGG16 and ResNet18 backbones with custom multi-layer classifiers, AdamW discriminative fine-tuning, and OneCycleLR scheduling for highly accurate food image categorization.
• animal-detection-autonomous-vehicles - Deep Learning - Computer vision classification pipeline using Convolutional Neural Networks (CNN) with batch normalization and global pooling to distinguish vehicles from animals for smart mobility safety.
• sentiment-analysis-MLOps - MLOps - Automated sentiment analysis and online reputation monitoring pipeline integrating the Twitter-RoBERTa model with a complete CI/CD and production-ready MLOps retraining workflow.
• exotic-fruit-recognition - Models and Algorithms - Automated fruit classification using the K-Nearest Neighbors (KNN) algorithm with hyperparameter optimization and cross-validation for high-accuracy quality control.
• real-estate-price-prediction - ML Fundamentals - Real estate price prediction using regression regularization to design a balanced model via Ridge, Lasso, and Elastic Net regressions with polynomial feature expansion and cross-validation.
• cpp-CRM - A production-ready C++ CRM engine utilizing object-oriented inheritance, hash-map constraints, and a fault-tolerant CSV persistence layer for real estate client and interaction management.