Security Researcher and AI Cybersecurity Engineer

I am a security researcher and Python engineer focused on vulnerability discovery, AI-native security tooling, and adversarial robustness. I build practical systems that help identify root-cause vulnerabilities across large and complex attack surfaces.

I architect and lead NIGHTSHADE, an AI-powered cyber operations platform that combines LLM orchestration, structured reasoning, and security automation. My work includes bespoke AST parsers, fuzzers, CVSS evaluators, and asynchronous analysis pipelines.

I have conducted independent security research with confirmed findings affecting Netflix, Circle (USDC/CCTP), and Lightspark. My research experience spans ReDoS, authorization and authentication bypasses, rate-limit failures, denial-of-service risks, and smart-contract security.

I am experienced in applied machine learning and AI security, including PyTorch, TensorFlow, transformers, RAG systems, NLP, adversarial ML, and LLM red teaming. I also developed medical AI research using Swin Transformer and DenseNet-169 for bone-fracture detection and adversarial robustness evaluation.

My engineering background includes Python, Flask, REST APIs, PostgreSQL, Docker, Linux, CI/CD, and cloud platforms. I hold CompTIA CySA+ and PenTest+ certifications and have achieved a Top 1% global ranking on TryHackMe.

I am based in Dhaka, Bangladesh, and am focused on cybersecurity engineering, security research, AI security, and security automation opportunities.