Projects

Selected projects across reservoir engineering, software, and machine learning — full details inline.

image:
assets/img/projects/slugguard.png

SlugGuard — Expert System for AI-Driven Slugging Detection in Wells

Reservoir EngineeringSoftware Engineering

A complete framework for detecting slugging in producing wells. The system includes the full end-to-end architecture: a module that mimics live sensor data, a database to store features and labels, a rule-based decision agent, and a notification system that alerts operators when slugging conditions are detected.

PythonExpertaSQL
image:
assets/img/projects/overfit.png

Overfit Potential of ANNs and Its Diagnosis

Machine LearningStatistics

Demonstrated the susceptibility of artificial neural networks to overfitting and designed multiple diagnostic and model-checking strategies to evaluate true generalization beyond the conventional validation-metric view. The project explores where standard validation hides overfitting and how to surface it.

PythonPyTorchTensorFlowOpenCV
image:
assets/img/projects/epattern.png

ePattern — Customizable Reservoir Visualization Toolkit

Software EngineeringReservoir Engineering

A web-based, customizable visualization tool developed at ei-Link R&D. ePattern lets engineers analyze data beyond standard visualization, evaluate reservoir health across areas, and run detailed diagnostics of key parameters. I architected and built the backend, and boosted performance by 220% through multithreading and caching.

C#SQLAngularREST APIsStatistics
image:
assets/img/projects/loan.png

Loan Default Prediction with ML — 1st Place

Machine LearningStatistics

Part of the ABB Innovation Center Machine Learning BootCamp. Predicted the probability of bank customers defaulting on loans using advanced gradient-boosting and ensemble techniques, with careful feature engineering and model evaluation — securing 1st place in the competition.

PythonScikit-learnXGBoostLightGBM