ayush dutta

I'm building an adaptive ML-based log anomaly detection system at MAKAUT using Isolation Forest and LogBERT to identify malicious patterns in real time.

I developed end-to-end log preprocessing, feature extraction, classification, terminal alerts, and a Flask dashboard for SOC-style monitoring. By applying NLP-based contextual analysis, I reduced false positives in anomaly classification while improving threat-detection recall through adaptive retraining.

During my cybersecurity internship with NIELIT Virtual Academy, I performed vulnerability analysis across TLS/SSL, HTTPS, defensive network architecture, threat modelling, and risk assessment using OWASP- and NIST-aligned practices.

I also built a Firebase-powered grocery inventory and billing platform that reduced manual data-entry and processing time by 40% through automated stock tracking, real-time updates, authentication, and role-based access control.