I am a Data Scientist and Machine Learning Consultant specializing in predictive modeling, customer segmentation, and custom BI solutions.
Leverage historical data to predict future trends. I build robust machine learning models (Random Forests, Gradient Boosting, GANs) to forecast market movements, customer churn, and risk defaults.
Maximize marketing ROI by understanding your audience. I use unsupervised clustering and Natural Language Processing (LDA, Sentiment Analysis) to turn raw customer data into actionable personas.
Stop guessing and start tracking. I design automated, interactive dashboards using Tableau, Power BI, and Python to give executives real-time visibility into critical KPIs.
Client: Aiolux
The Challenge: Predict whether 11 S&P 500 sectors would outperform the market based on CPI, despite having a highly constrained, small historical dataset.
The Solution: Engineered Tabular Generative Adversarial Networks (TGAN) to generate synthetic economic data. Applied SMOTE and deployed 33 Extremely Randomized Tree Classifiers to achieve 78.8% accuracy over 200-day horizons.
Client: Kearny Bank
The Challenge: Generic marketing campaigns were resulting in low conversion rates. The bank needed actionable, data-driven customer personas.
The Solution: Led unsupervised clustering (K-means, DBSCAN, PCA) on transaction behavior. Integrated the clusters with NLP topic modeling from customer reviews to design highly targeted A/B tested campaigns.
Client: Kearny Bank
The Challenge: Traditional supervised models were failing to detect high-risk loans due to extreme class imbalance (defaults were rare).
The Solution: Reframed the problem as anomaly detection. Built and deployed Isolation Forests and One-Class SVMs to flag deviations in credit behavior, enabling underwriters to intervene early.
Collaborated with product managers and data engineers to align analytics with business priorities. Built anomaly detection models to flag high-risk loans and led customer segmentation initiatives that directly increased campaign ROI by 30%. Designed executive dashboards in Tableau and Power BI.
Built a real-time BMI prediction pipeline via a VGG-16 Convolutional Neural Network (CNN). Deployed the machine learning model into production using a Flask API integrated with live video input feeds.
Focused on financial forecasting and model deployment at scale. Identified macroeconomic drivers using causal inference and utilized Generative Adversarial Networks (GANs) to generate synthetic data, improving overall model robustness by 20%.