Building models that actually generalise.

Developed a classification model designed to identify Pokémon in complex environments. The project consisted in Feature engineering and data augmentation, utilized Multi-Layer Perceptrons (MLPs) and then Convolutional Neural Networks (CNNs) to optimize accuracy and model generalization.

Conducted predictive modeling to estimate survival timelines for patients diagnosed with Multiple Myeloma. Applied dimensionality reduction techniques like Isomap and non-parametric algorithms such as K-Nearest Neighbors (KNN). Analyzed high-dimensional clinical data to identify key prognostic factors and enhance survival probability forecasting.

Developed a comprehensive Big Data ecosystem to analyze and monitor 173 million taxi rides in New York City. Engineered a dual-layer pipeline that transitions from high-scale batch processing—using grid-mapping and spatial-temporal analysis to identify revenue hotspots—to a live streaming solution that processes incoming ride data in real-time, providing immediate actionable insights into urban mobility patterns.

Engineered a machine learning framework to predict chaotic satellite trajectories in the Three-Body Problem using KNN Regression and polynomial models. Optimized the predictive accuracy by conducting rigorous hyperparameter tuning for K and evaluating performance across metric spaces.

Conducted a deep-dive into Combinatorial Game Theory and Economic models, developing Python algorithms to calculate optimal winning strategies. Implemented an object-oriented simulation.

Designed and deployed a nationwide questionnaire to analyze the correlation between mental/physical health and workplace productivity.

Developed an application to automate charcoal production workflows.

Led the end-to-end modeling and development of a digitized sports management platform.

Developed a clinical data analysis to evaluate the efficacy of a novel cholesterol treatment.