model-serialization

Pickle-free model bundles: one verifiable file with the model, a rich manifest, schema and structure for scikit-learn, XGBoost, CatBoost and ONNX.

Serialization and Model serialization.

The aim of this project is to predict the sales six weeks ahead across all the stores of the Rossman Pharmaceutical company using Machine and Deep Learning. The different factors affecting the sales are: promotions, competitions, school-state holiday, seasonality, and locality.

Full-stack Flask insurance workflow system with policy management, QR verification, Twilio SMS alerts, NLP sentiment analysis, and Decision Tree-based fraud detection (77% accuracy).

IMDB API Clone which is used to review and add the movie ratings

A diagnostic neural network built in Python to classify diabetic risk based on clinical parameters, optimized using Batch Normalization and ReLU, and serialized for production use.

Robust CNN & CapsuleNet benchmarks for noisy images – modern TensorFlow 2/Keras 3 support.

A fully modular end-to-end Machine Learning pipeline for Iris classification. Includes data preparation, scaling, model training, evaluation, performance comparison, and a CLI-based prediction system using Logistic Regression, Decision Tree, and SVC. Designed as a clean template for beginners to understand core ML workflow.

Sales forecasting and customer behavior analysis for Rossmann Pharmaceuticals using machine learning and deep learning techniques.

Tech Challenge of the Postgraduate in Data Analytics, from FIAP, analyzing Brent Oil price data, in comparison with historical, economic and societal data, integrating correlation and causality analyzes of items with prices, as well as developing a model forecast and an importance analysis through information gain from a forest model (XGBoost)

Machine learning project for classifying parking spaces as occupied or empty using Support Vector Machines (SVM).

Streamlit university-admission prediction learning project using linear regression, feature scaling, and serialized model artifacts.