Projects

Hybrid log classification system combining regex-based parsing with LLM-powered analysis for intelligent log categorization and anomaly detection.

End-to-end machine learning pipeline for telecom customer churn prediction with model training, evaluation, and a production-ready API.

A transformer-based GPT model built from scratch using PyTorch, implementing attention mechanisms, positional encoding, and text generation.

Convolutional neural network pipeline for image classification with data augmentation, training, and evaluation.

Unsupervised machine learning system for customer segmentation using clustering algorithms to identify distinct customer groups.

Data analytics project analyzing HR and recruitment datasets, building dashboards, and delivering actionable insights for hiring decisions.