End-to-end machine learning system for predicting vehicle insurance policy lapse.
End-to-end machine learning system to identify customers at risk of policy lapse.
Built using behavioural and policy data with feature engineering and model optimisation.
Task: Binary Classification
Projects focused on model development, training, and evaluation.
Built a supervised learning model to identify high-income individuals likely to donate.
Task: Binary Classification
Python • Scikit-learn • Pandas
MobileNetV2-based image classification using transfer learning on the Oxford Flowers 102 dataset.
Task: Multi-class classification (102 categories)
Python • PyTorch • MobileNetV2 • Transfer Learning
Projects focused on data analysis, insights, and decision-making.
Analyzed smart device usage data to identify behavioral patterns and generate actionable insights.
Focus: Behavioral Data Analysis
SQL • Python • Data Analysis • Visualization
Explored user behavior trends to support data-driven marketing strategies and business decisions.
Focus: Exploratory Data Analysis
Python • Excel • Data Analysis • Visualization
Experimental work focused on dynamical systems and optimization techniques.
Implemented the Next Generation Reservoir Computing framework to model chaotic dynamical systems (Lorenz-63, Double Scroll), achieving accurate short-term predictions and capturing system dynamics.
Focus: Dynamical Systems & Nonlinear Modeling
Python • Dynamical Systems • Reservoir Computing • Time Series
Implemented and evaluated clustering algorithms (k-means, hierarchical, and density-based), identifying limitations of k-means for non-spherical data and comparing standard and mini-batch variants.
Focus: Unsupervised Learning & Clustering
Python • Scikit-learn • Clustering • Unsupervised Learning
Analyzed the Adam optimization algorithm, studying convergence behavior and the impact of adaptive learning rates and hyperparameter tuning.
Focus: Optimization & Machine Learning Theory
Optimization • Gradient Descent • Machine Learning