Shourya Negi

At GeeksforGeeks, I analysed organisational datasets with SQL and Python to investigate trends, anomalies, and data-quality issues. I also helped automate recurring reporting steps to support more consistent analysis.

For my MSc dissertation, I built and compared approaches to predict 30-day hospital readmission risk, including XGBoost, Random Forest, and Logistic Regression. I used SHAP to explain predictions and communicate their limitations.

On a university team project, I led a four-person effort modelling metal-parts lifespan and safety. Random Forest lifespan prediction achieved an R² of 0.94, while Gradient Boosting safety classification achieved 97% accuracy.

I also created an employee behaviour analytics dashboard using MySQL and Power BI, analysing 10,000 survey records. I’m motivated to develop in insurance, pensions, and actuarial consulting through EY’s 2027 programme.