published work

Research Papers

Two papers born from long nights, longer literature reviews, and a genuine love of finding patterns in messy, real-world data.

2025
Accepted for Publication

AI-Driven Drug Discovery for Dengue using Ensemble, Graph Neural Networks, and Generative Models

Designed an AI-driven framework combining Random Forest, XGBoost, Graph Neural Networks (GCNs), and Variational Autoencoders (VAEs) to predict dengue drug efficacy from molecular descriptor data. The GNN model delivered the strongest performance at 91.2% accuracy, ahead of XGBoost (89.14%) and Random Forest (86.29%), with results validated using Precision, Recall, F1-score, and ROC-AUC.

Machine LearningGNNXGBoostFeature EngineeringVAEs
2026
Manuscript in Publication Process

Drift-Aware Hybrid Machine Learning and Deep Learning Framework for Predictive Anomaly Detection and Safe Recovery Recommendation in Cloud Infrastructure Systems

Proposed a hybrid ML/DL framework combining drift detection with predictive anomaly detection to enable safe, automated recovery recommendations for cloud infrastructure systems — aimed at improving system reliability and reducing downtime in dynamic cloud environments.

Machine LearningDeep LearningTime-Series AnalysisAIOpsXGBoostLSTM AutoencodersDrift DetectionCloud AnalyticsPredictive MonitoringAnomaly Detection