Cybersecurity Researcher at Dayananda Sagar Academy of Technology and Management and Volunteer at AI-Sec Community.
Speaker
Cybersecurity Researcher at Dayananda Sagar Academy of Technology and Management and Volunteer at AI-Sec Community
Taha Kathiria is a cybersecurity researcher and enthusiast with a strong focus on AI-driven threat detection, vulnerability assessment, and SOC automation. His work includes published research in IEEE conferences, particularly on dynamic ransomware detection in MySQL databases. As a Cyber Community Intern at Precogly.org, Taha is actively involved in fostering cybersecurity awareness and education. With expertise in Python, machine learning, and SIEM tools, he is committed to advancing cybersecurity through innovation and community engagement.
Database ransomware is a growing threat, exploiting legitimate SQL credentials to evade traditional defenses. This talk introduces DB Sherlock, a multi-modal MySQL detection system combining static signatures, neural sequence models, and real-time entropy analysis. Trained on 50,000 hours of benign activity and 5,000 attack cases, DB Sherlock achieves 97.2% detection, 1.8% false positives, and sub-9-second response times with minimal overhead. The presentation will highlight its architecture, live detection capabilities, and how AI-driven behavioral profiling can transform database security against evolving ransomware attacks.