Aarush User

Second-Year Computer Science with Artificial Intelligence student at the University of Nottingham Malaysia, with research experience in machine learning and hands-on software development.

As part of my 2026 research internship, I benchmarked 8 forecasting methods across 4 public network-traffic datasets using chronological train-test evaluation. LSTM+GRU achieved R² = 0.9634 on AIviVN ZONE01 with approximately 56% lower RMSE than the Persistence baseline, while the broader comparison showed that no single model family performed best across every dataset.

The project strengthened my interest in model evaluation, time-series forecasting and applied machine learning, particularly the importance of testing complex methods against simple, well-defined baselines.

Outside ML research, I’ve built a full-stack PHP/MySQL internship-assessment system and a concurrent TCP client-server suite in C using select() and multithreading. I primarily work with Python, TensorFlow, scikit-learn, NumPy, pandas, C, SQL and Git.