Sidharth M

BTech — Computer Science & Engineering (Cybersecurity)

Chennai, Tamil Nadu, India

Kozhikode, Kerala, India · Government Cybersecurity Division

• Continued contributions to cybersecurity threat analysis, community security awareness programs, and operational security support.

• Performed phishing campaign analysis — mapped attacker TTPs, identified malicious infrastructure, and contributed to threat detection improvements.
• Deployed and configured OPNsense next-generation firewalls; documented firewall rule bypass techniques for internal red-team research.
• Supported red teaming exercises and network reconnaissance to identify security gaps.

VIT Chennai, Tamil Nadu, India

• Elected Vice Chair; lead chapter strategy, manage website security posture, and oversee technical operations for a 200+ member student tech community.

• Managed and hardened the ACM VIT Chennai website; led technical event operations and membership onboarding.

Research Intern

Supervisor: Dr. T. Subbulakshmi · VIT Chennai

• Built a comprehensive cloud-native attack dataset and ML categorization model to classify cloud security threats.
• Bridged threat intelligence with machine learning; research culminated in a peer-reviewed IJCNC publication.

Subbulakshmi T, Adithyan P, Sidharth M, Arun Santhosh R A

International Journal of Computer Networks & Communications (IJCNC), Vol. 18, No. 2 · aircconline.com

T. Subbulakshmi, Aditya Kushwaha, Sidharth M

Open Source For You (OSFY) Magazine · opensourceforu.com

Brindha Subburaj, Subbulakshmi T, Sidharth M

IGI Global · igi-global.com

• In-depth technical exploration of NGFW architectures, deep packet inspection, and enterprise deployment strategies.

Docs: error-siddh.github.io

• Built EchoVault, a full-stack real-time transcription system that captures live microphone audio, performs on-device AI transcription, and streams results to a collaborative web dashboard.
• Engineered a modular Python + FastAPI backend with WebSocket telemetry (mic levels, CPU/RAM/GPU stats, model loading state), robust audio resampling pipeline, and REST-based control plane for production-style operability.
• Integrated multi-model AI pipeline: faster-whisper for speech-to-text, ECAPA-TDNN for speaker footprinting, and Qwen 0.5B for offline meeting summarization with post-session assistant workflow.
• Developed a modern Next.js 15 + React 19 frontend with real-time transcript feed (WebSockets/SSE), room-based live updates, and MongoDB-backed stateless memory to preserve transcript context across backend restarts.
• Technologies: Python, FastAPI, faster-whisper, ECAPA-TDNN, Qwen, WebSockets, SSE, Next.js, React, Tailwind CSS, ShadCN, MongoDB, Linux, NVIDIA GPU telemetry (nvidia-smi), psutil, librosa.

Docs: error-siddh.github.io

• Modular Bash toolkit implementing CIS Ubuntu 22.04 LTS Level 1 hardening across authentication, SSH, filesystem, kernel/network sysctl, logging, firewall (UFW + iptables), AppArmor, and sudo controls.
• Firewall module implements defence-in-depth: SSH brute-force rate limiting, anti-spoofing (RFC 1918 drops), scan protection (NULL/XMAS/FIN/non-SYN drops), ICMP hardening, and strict outbound allowlist.
• Compliance validator runs 73 CIS controls, generating PASS/FAIL/WARN HTML dashboard — achieved 73% compliance score (54 PASS) with auto-rollback on failure.
• Technologies: Bash, UFW, iptables, auditd, rsyslog, AppArmor, Linux (Ubuntu 22.04).

GitHub: github.com/ERROR-SIDDH/AEGIS · Docs: error-siddh.github.io/AEGIS/ · Live: aegis-gules.vercel.app

• Production-ready, open-source examination management platform built on Next.js 14 + MongoDB; deployed via Docker on Vercel.
• Enforced 5-point server-side validation chain: PC approval → student mapping → exam assignment → status check → duplicate submission guard.
• AI-powered question bank using Google Genkit (Gemini) — auto-classifies difficulty, generates semantic tags, supports LaTeX and code rendering.
• Real-time PC telemetry: heartbeat polling every 15 seconds with liveStatus tracking (Online → Attempting → Finished); full admin audit logging.
• Technologies: Next.js 14, TypeScript, MongoDB, Google Genkit, Docker, Tailwind CSS, Vercel.

Docs: error-siddh.github.io

• Conducted comprehensive forensic analysis of a volumetric UDP flood attack (152k+ PPS) targeting a FiveM gameserver.
• Stream-parsed and analyzed over 18.3 million packets from raw binary PCAP data using Python (dpkt).
• Developed an automated data pipeline to compute traffic metrics, detect IP spoofing through TTL analysis, and perform geo-enrichment via MaxMind GeoLite2.
• Created interactive visual synthesis using Plotly and Folium to map global botnet distribution, attack origin clusters, and temporal protocol shifts.
• Technologies: Python, PCAP analysis, dpkt, MaxMind GeoLite2, Plotly, Folium.