Research - Antony Shane, Adversarial AI, Malware Engineering, Vulnerability Research

Cybersecurity Research.

Focused on adversarial AI, adaptive malware engineering, and vulnerability research under India's National Security Council Secretariat. Targeting SCI-indexed journals and leading IEEE security conferences.

Adversarial AI Malware Research Vulnerability Research IEEE / SCI Offensive Security

Current work.

In Progress · 2026

Adaptive Malware Evolution Framework for EDR/XDR Resilience Evaluation

An adaptive malware framework where a persistent implant on an EDR-protected host exfiltrates telemetry to a C2 server, which replicates the target environment in VMs running the same EDR to serve as a black-box optimization oracle. A hybrid approach combining memetic algorithms for behavioral strategy search and GNN-based CFG mutation for structural variation evolves command execution sequences that evade detection, converging within the window between EDR content updates. The core contribution is demonstrating that malware can autonomously discover and exploit blind spots in ML-based behavioral detection through closed-loop optimization — formally defining the behavioral mutation space, characterizing the sim-to-real transfer gap between lab VMs and live targets, measuring re-evolution cost after EDR updates, and evaluating cross-vendor transferability to determine whether evasion variants exploit fundamental weaknesses in behavioral ML or product-specific gaps.

Python Adversarial ML Memetic Algorithms GNN / CFG Mutation EDR/XDR Behavioral Analysis Closed-Loop Optimization

↳ Targeting SCI-indexed journals & IEEE S&P / USENIX Security

Ongoing · 2024–Present

Open-Source Vulnerability Research (NSCS Program)

Identification of critical vulnerabilities and exploit chains in open-source software ecosystems as part of a cybersecurity research project under India's National Security Council Secretariat (NSCS). Includes responsible disclosure, adversarial modeling simulation, and AI-driven defense system development. Advised by Dr. Durga Prasad Mohapatra, NIT Rourkela.

Exploit Research NSCS India Open-Source CVE Disclosure Adversarial Modeling

Domains & focus areas.

Adversarial AI & ML Security

Evasion attacks against ML-based detection, adversarial example generation, and robustness evaluation of security-critical models. Includes LLM-guided attack automation.

Malware Engineering & Evolution

Dynamic code mutation, behavioral transformation, anti-analysis techniques, and evolutionary algorithm-driven evasion against modern EDR/XDR pipelines.

Vulnerability Research & Exploit Development

Systematic discovery of critical vulnerabilities and exploit chains in open-source software. Coordinated disclosure with national security relevance under NSCS India.

AI-Driven Offensive Security

Autonomous penetration testing frameworks using LLM agents, feedback-driven attack loops, and self-refining exploitation strategies across web, network, and binary targets.

Digital Forensics & Incident Response

Post-incident forensic analysis on live ransomware and APT scenarios. Evidence acquisition, timeline reconstruction, and AI-assisted artifact correlation.

Threat Detection & AI Defense

Ensemble ML architectures (Random Forest, XGBoost, LSTM, CNN) for real-time threat detection. MITRE ATT&CK-aligned automated response and SOC pipeline optimization.

Research partnerships welcome.

Open to research collaborations across offensive and defensive security, adversarial AI, and intelligent defense systems. If your work involves breaking, building, or hardening security — let's talk.