Ishan Panwar
About
M.Tech student in Cybersecurity at Delhi Technological University, researching defenses against backdoored AI models. He previously completed a B.E. in Electronics and Computer Engineering at Thapar Institute of Engineering and Technology and has worked on security and cloud scanning at iTRUSTXForce and as a trainee at Bharat Electronics. He coauthored a 2026 ICSCCC conference paper on detecting backdoored models through gradient-entropy analysis and is a co-inventor on a published Indian patent application concerning adaptive routing for underwater software-defined networks.
Research focus · AI Backdoor Defense
Key Projects
AI Backdoor Defense
studies model behavior and gradient-entropy signals for detecting backdoored models from adversarial inputs.
NES2ELF Static Recompiler
explored binary translation and compiler tooling using Rust and LLVM.
Linux Li-Fi Drivers
experimental Raspberry Pi/Linux driver project for light-based data communication, published as Laser-Driver.
PS2Translate
experimental PlayStation 2 binary translator built with Rust and LLVM tooling.
More projects (1)
Trust-Aware Underwater Routing
contributed to a Q-learning-based routing system for underwater software-defined networks.
Patents
- S. Singh, A. Verma, I. Panwar, P. Kaushal, D. Gupta and P. Chawla, "A Real-Time Adaptive Trust Routing for Underwater Software-Defined Networks System," Indian Patent Application No. 202511061850, filed June 28, 2025 and published July 11, 2025.
Publications · 1
- Conference paperDetecting Backdoored Models using Gradient Entropy Analysis of Adversarial InputsIshan Panwar, Anukriti Kaushal, Abhishek Jain, Rajeev Kumar, Altaf Husain, Jitendra Singh2026 Fourth International Conference on Secure Cyber Computing and Communications (ICSCCC), pp. 1007–1012 · 2026DOI: 10.1109/ICSCCC69031.2026.11600379 ↗