FairClaim — Medicare Fraud Detection

Terno AI | 2 weeks | May 2026

Machine learning and explainable AI for detecting potentially fraudulent Medicare claims.

• Built Logistic Regression, Random Forest, and XGBoost models

• Engineered claim-level features from Medicare datasets

• Developed a deployment-ready fraud detection application with SHAP explainability

Clinical Trial Patient Attrition Prediction

Terno AI | 2 weeks | May 2026

Predictive analytics to identify patients at higher risk of dropping out of clinical trials.

• Developed and evaluated multiple ML models for dropout prediction

• Identified key challenges in model generalisability across datasets

• Designed a standardized ML framework for scalable clinical trial analysis

Global Mental Health Analytics

Terno AI | 2 weeks | April 2026

Multi-modal analytics combining social media, sentiment, and global suicide statistics to identify mental health risk patterns.

• Analysed 20,000+ data points

• Developed ML pipelines for risk scoring and cohort segmentation

• Identified behavioural patterns associated with higher-risk groups

Urban Risk Analytics for Delivery Worker Safety

Terno AI | 2 weeks | April 2026

Data-driven risk prediction using delivery, weather, customer, and road-safety data.

• Integrated multiple data sources for urban risk analysis

• Developed Random Forest and XGBoost risk models

• Explored applications in safer routing, timing, and workforce safety

Pixel2Number — Image Recognition

MIT | September 2025

Deep learning for recognising digits in real-world street images using the SVHN dataset.

• Compared Feedforward and Convolutional Neural Networks

• CNN achieved the strongest performance

• Built using Python, TensorFlow/Keras, NumPy, and Matplotlib

FoodHub — Customer & Business Analytics

MIT | August 2025

Exploratory and statistical analysis of food-ordering data to uncover customer and business insights.

• Analysed ordering behaviour and restaurant performance

• Identified trends, patterns, and anomalies

• Created visualisations and data-driven business insights

From healthcare and biomedical analytics to AI, machine learning, and predictive modelling, these projects demonstrate our ability to turn complex data into practical, decision-ready solutions.