# Rumman Firdos — Full Site Content for AI > **Cybersecurity researcher, AI engineer, and technology entrepreneur.** Building at the intersection of cybersecurity, artificial intelligence, and technology entrepreneurship. > > **Website:** https://rummanfirdos.com > **Contact:** hello@rummanfirdos.com > **GitHub:** https://github.com/Arnosdeus > **Google Scholar:** https://scholar.google.com/citations?user=sdEqPUsAAAAJ&hl=en > **LinkedIn:** https://www.linkedin.com/in/rumman-firdos > **ORCID:** https://orcid.org/0009-0007-4021-650X --- ## About Rumman Firdos Rumman Firdos is a cybersecurity researcher, AI engineer, and technology entrepreneur. His work spans adversarial machine learning, autonomous security operations, detection engineering, and AI security. He founded Scalewidth (autonomous detection infrastructure), Brevton Group (technology and AI), and Exertus (open-source security tools). ### What I'm Working On 1. **SecureScan** — AI-powered security analysis platform combining heuristic analysis, ML classification, and VirusTotal validation for URL, file, and hash analysis. Published platform release (2025). 2. **ComputeChain** — Decentralized computing architecture research for intelligent scheduling and distributed resource coordination. Currently under submission to Springer LNNS (2026). 3. **Vulnerability Shift in NIDS** — Adversarial ML research on multi-epsilon curriculum learning for intrusion detection systems. Submitted to Springer LNNS (2026). 4. **Scalewidth** — Leading the engineering and research behind autonomous detection infrastructure for enterprise security operations. 5. **Enterprise security platforms** — Building practical enterprise cybersecurity products and AI-powered systems, including Vigilante (endpoint protection) and Enterprise Portal. ### Professional Timeline - **2025** — Published SecureScan, the AI-powered security analysis platform. - **2024** — Founded Scalewidth (autonomous detection infrastructure) and Brevton Group (technology and AI). - **2023** — Founder of Vigilante, the Scalewidth cybersecurity division. Led early work on autonomous endpoint detection and self-healing detection rule systems. - **2021** — Security engineer in detection engineering. Built detection-as-code pipelines and security automation at scale. ### Professional Profiles | Platform | URL | |----------|-----| | GitHub | https://github.com/Arnosdeus | | Google Scholar | https://scholar.google.com/citations?user=sdEqPUsAAAAJ&hl=en | | arXiv | https://arxiv.org/search/?searchtype=author&query=Firdos%2C+R | | Semantic Scholar | https://www.semanticscholar.org/author/Rumman-Firdos/2410030889 | | ORCID | https://orcid.org/0009-0007-4021-650X | | ResearchGate | https://www.researchgate.net/profile/Rumman-Firdos | | LinkedIn | https://www.linkedin.com/in/rumman-firdos | | Wikidata | https://www.wikidata.org/wiki/Q140782484 | --- ## Research Papers ### SecureScan: AI-Powered Security Analysis Platform **Authors:** Rumman Firdos **Venue:** Platform Release **Year:** 2025 **Status:** Published **URL:** https://rummanfirdos.com/research/securescan SecureScan is an AI-powered security scanner that combines heuristic analysis, machine learning classification, and VirusTotal validation for comprehensive URL, file, and hash analysis. The platform provides automated threat intelligence gathering and analysis for security practitioners. --- ### ComputeChain: A Decentralized Computing Architecture for Intelligent Resource Coordination **Authors:** Rumman Firdos **Venue:** Springer Lecture Notes in Networks and Systems (LNNS) **Year:** 2026 **Status:** Under Submission **URL:** https://rummanfirdos.com/research/compute-chain ComputeChain proposes a decentralized computing architecture designed to coordinate distributed computing resources through intelligent scheduling, scalability, and decentralized infrastructure. The architecture addresses key challenges in resource allocation, fault tolerance, and coordination across heterogeneous computing environments. --- ### Mitigating Vulnerability Shift in Adversarially Hardened Network Intrusion Detection Systems via Multi-Epsilon Curriculum Learning **Authors:** Rumman Firdos **Venue:** Springer Lecture Notes in Networks and Systems (LNNS) **Year:** 2026 **Status:** Submitted **URL:** https://rummanfirdos.com/research/vulnerability-shift-nids Introduces curriculum-based adversarial training to mitigate vulnerability shift within adversarially trained intrusion detection systems while improving robustness against multiple adversarial attacks. The approach uses multi-epsilon curriculum learning to progressively harden models against increasingly sophisticated attacks, addressing the defense budget exhaustion problem. --- ## Technical Writing ### Detection Engineering at Scale: A Framework for Autonomous Security Operations **URL:** https://rummanfirdos.com/writing/detection-engineering-at-scale **Date:** June 2026 **Reading time:** 14 min read How to design detection pipelines that learn from signal, reduce analyst fatigue, and operate autonomously — without adding noise. Covers feedback loops, signal-to-noise metrics, automated tuning, and real-world deployment at Scalewidth. ### Adversarial Robustness of Machine Learning Systems in Production **URL:** https://rummanfirdos.com/writing/adversarial-robustness-llm-production **Date:** April 2026 **Reading time:** 11 min read A practical examination of adversarial attack surfaces in deployed ML systems and mitigations that hold up at runtime. Covers evasion, extraction, inversion, and prompt injection from an engineering perspective. ### The Emerging Threat Landscape of LLM-Powered Applications **URL:** https://rummanfirdos.com/writing/llm-security-landscape **Date:** February 2026 **Reading time:** 17 min read Prompt injection, model inversion, data extraction — mapping the attack surface of applications built on large language models. Includes real-world case studies and a defense framework. ### Building Security Automation That Does Not Add Noise **URL:** https://rummanfirdos.com/writing/security-automation-noise **Date:** November 2025 **Reading time:** 9 min read Why most automation in security operations increases alert volume, and how to design systems that reduce it. Covers feedback-driven threshold tuning, signal measurement, and practical implementation patterns. --- ## Projects ### Vigilante — Enterprise Endpoint Protection **URL:** https://rummanfirdos.com/projects/vigilante **Venture:** Scalewidth Enterprise endpoint protection and security platform designed to provide malware protection, endpoint visibility, enterprise management, policy enforcement, and intelligent security operations. **Technologies:** Python, Flask, React, MongoDB, AI ### Enterprise Portal **URL:** https://rummanfirdos.com/projects/enterprise-portal **Venture:** Scalewidth Enterprise security platform for comprehensive security operations management. **Technologies:** Python, React, Node.js ### SecureScan **URL:** https://rummanfirdos.com/projects/securescan **Venture:** Brevton Group AI-powered malware, phishing, URL, and file analysis platform that combines heuristic analysis, machine learning, and external threat intelligence for comprehensive security analysis. **Technologies:** Python, Flask, Machine Learning, React ### StegnoCrypt **URL:** https://rummanfirdos.com/projects/stegnocrypt Steganography and cryptography tool for secure data hiding and encryption. ### ComputeChain Platform **URL:** https://rummanfirdos.com/projects/computechain-platform Distributed computing platform inspired by the ComputeChain research initiative. Enables intelligent scheduling and coordination of computing resources across decentralized infrastructure. **Technologies:** Python, Docker, Distributed Systems ### Autonomous SOC **URL:** https://rummanfirdos.com/projects/autonomous-soc **Venture:** Scalewidth Multi-agent autonomous Security Operations Center for intelligent detection, investigation, and response. Leverages AI agents to automate security operations workflows and accelerate incident response. **Technologies:** Python, AI Agents, LLM Applications --- ## Ventures & Organizations ### Scalewidth **URL:** https://rummanfirdos.com/ventures/scalewidth **Founded:** 2024 **Role:** Founder **Status:** Active **Industry:** Cybersecurity / Autonomous Security Leading the engineering and research behind autonomous detection infrastructure for enterprise security operations. Building detection pipelines that tune themselves from operational feedback. ### Brevton Group **URL:** https://rummanfirdos.com/ventures/brevton-group **Founded:** 2024 **Role:** Founder **Status:** Active **Industry:** Technology / AI A technology and AI company focused on bridging research and practical security applications. ### Exertus **URL:** https://rummanfirdos.com/ventures/exertus **Founded:** 2025 **Role:** Founder **Status:** Active **Industry:** Open-Source Security Tools An open-source engineering organization building cybersecurity tooling, AI technologies, developer tools, and research implementations. Focus on tools that practitioners can actually use. --- ## Media & Press **URL:** https://rummanfirdos.com/press Professional media resource for journalists, conference organizers, podcast hosts, universities, researchers, and industry publications. ### Biography - **50-word:** Rumman Firdos is a cybersecurity researcher, AI engineer, and technology entrepreneur focused on adversarial machine learning and autonomous security operations. He founded Scalewidth, Brevton Group, and Exertus. - **100-word:** Rumman Firdos is a cybersecurity researcher, AI engineer, and technology entrepreneur working at the intersection of adversarial machine learning, detection engineering, and autonomous security operations. He founded Scalewidth (autonomous detection infrastructure), Brevton Group (technology and AI), and Exertus (open-source security tools). His research focuses on how AI systems fail under adversarial pressure and on detection and response infrastructure that can keep pace with autonomous attackers. - **250-word:** Rumman Firdos is a cybersecurity researcher, AI engineer, and technology entrepreneur building at the intersection of adversarial machine learning, detection engineering, and autonomous security operations. Through Scalewidth, he develops autonomous detection infrastructure for enterprise security operations. Through Brevton Group, he works on technology and AI product development. Through Exertus, he supports open-source security tooling and research implementations. His research examines how deployed AI systems fail under adversarial input and how detection engineering can evolve into a fully autonomous discipline. He has published work on adversarial robustness, LLM security, and detection automation. ### Areas of Expertise Cybersecurity, Artificial Intelligence, Security Engineering, Autonomous Security Operations, Detection Engineering, Security Automation, Adversarial Machine Learning, AI Security, AI Agents, Technology Entrepreneurship ### Commentary Topics Artificial Intelligence, Cybersecurity, Security Engineering, Autonomous Security, AI Security, Enterprise Software, Technology Entrepreneurship, Security Research ### Media Contact - **Email:** hello@rummanfirdos.com - **Contact Form:** https://rummanfirdos.com/contact --- ## Current Focus (Now Page) **URL:** https://rummanfirdos.com/now Rumman is currently focused on: 1. **SecureScan** — AI-powered security analysis platform (published 2025) 2. **ComputeChain** — decentralized computing architecture research, under submission to Springer LNNS 3. **Vulnerability Shift in NIDS** — adversarial ML research on multi-epsilon curriculum learning, submitted to Springer LNNS 4. **Scalewidth** — leading engineering and research behind autonomous detection infrastructure 5. **Enterprise security platforms** — Vigilante (endpoint protection) and Enterprise Portal **Reading:** "Adversarial Machine Learning" by Biggio & Roli; "Designing Data-Intensive Applications" by Kleppmann; NIST AI Risk Management Framework. --- *This file is an AI-friendly full-site summary. For the complete experience, visit https://rummanfirdos.com*