I'm Zubair Usman — I help startups and growing businesses identify security risks, resolve technical problems, and automate repetitive workflows through reliable software engineering.
A snapshot of the work: what I've found, what I've shipped, and where I am right now.
Severity, reproduction steps and remediation notes are documented per authorised engagement. Specific findings stay private unless approved for publication.
Responsible disclosure and client confidentiality come first.
Ürümqi, China. Overlapping Pakistan, the Gulf and the EU daily, with US mornings on request.
OAuth 2.0 nonce flaw disclosed — Windsurf IDEReplay, CSRF and session-fixation vectors all confirmed
Bug-bounty recon — inDrive, Bugcrowd programsAttack-surface mapping across Chime and TheFork
Full web security auditDelivered with a fix roadmap and follow-up guidance
MSc research begins — Xinjiang UniversityIndustrial AI and anomaly detection
Current availability, rates and platform metrics are confirmed directly for each engagement.
Security audits, ISO 27001 gap analysis, an enterprise ERP build and bug-bounty reconnaissance.
Start with a bounded problem. Pricing shown here is a proposed starting point; final scope, access and deliverables are confirmed in writing.
Authorized testing and evidence-led reporting for one application or API.
Turn repetitive operations, data work and API handoffs into dependable workflows.
Focused troubleshooting for plugin conflicts, performance and configuration issues.
Explore real-world software, cybersecurity tools, and engineering projects.
Selected systems and research builds, from a production VPN platform to open-source security tooling. Each case study separates public proof from owner-review items.
OpenVPN + WireGuard server automation with a customer portal, diagnostics and metrics — running on Alibaba Cloud ECS behind nginx.
vpn.zus3c.infoSales and purchase platform with AI-agent RFQ processing, barcode-verified PDFs and role-based approval workflows.
Eight department agents — research, lead-gen, bug-bounty scoping, engineering — each gated by its own permission profile, with a strict no-offensive-automation rule.
One script that installs and configures 300+ offensive-security tools across 20+ categories, with parallel installs, dependency management and isolated Python environments.
github.com/offsec-toolkitStatic malware detection on PE/ELF headers, entropy and API imports using Random Forest and XGBoost, with SHAP explainability and a hardened FastAPI service.
github.com/MalwareGuard-AIA local-first AI coding assistant built on Ollama: modular agents for code analysis, debugging and safe file editing — fully offline, nothing leaves the machine.
github.com/LocalCodex-AITurns large, messy scan JSON into an actionable security report — custom API integration and automated reporting in under three minutes.
github.com/vulnerability-scan-parserNetwork security scanner with device discovery, packet analysis and alerts written in plain language a non-engineer can act on.
A bug-bounty learning platform: a full roadmap from reconnaissance to web and mobile testing, built for beginners who want a real path into offensive security.
github.com/BugDrillXTurns any webcam into a hands-free cursor using facial gestures — built for people with mobility challenges, powered by computer vision and machine learning.
github.com/Smart-Cursor-System-2.0Browser-automation suite for WhatsApp Web: CSV upload, QR authentication, session persistence and a clean dashboard for campaign sends.
Public repositories on GitHub — security tooling, ML experiments, cheat sheets and teaching material.
Installs and configures 300+ offensive-security tools across 20+ categories — parallel installs, GitHub API integration, isolated Python environments.
AI-powered static malware detection on PE/ELF binaries with Random Forest / XGBoost, SHAP explainability and a FastAPI deployment.
Local-first AI coding assistant on Ollama: multi-model routing, agent system, memory and safe file edits — offline and private.
A complete API security audit framework: checklists, testing methodology and compliance guidance for auth, data exposure and rate limiting.
Parses large-scale vulnerability scan output and turns complex JSON logs into an actionable report in under three minutes.
A working reference on Server-Side Request Forgery in cloud-native and AI-driven architectures, and how modern SSRF bypasses defences.
A bug-bounty education platform with a full roadmap from reconnaissance to web and mobile testing. Powered by SCOLTECH.
Accessibility tool that turns a webcam into a hands-free cursor using facial gestures, computer vision and machine learning.
A curated, free collection of ready-to-use AI prompts for developers, security professionals, SaaS builders and creators.
A desktop Markdown editor with live HTML preview, split-pane interface, formatting toolbar and export — built with Python and Tkinter.
Packet Tracer environments for simulating smart-factory networks — a teaching resource for researchers and students.
Core .NET library for a lighting management information system: business logic, data access and shared functionality.
Every finding reproducible, CVSS-scored and reported responsibly. Tap a row to open it.
A missing nonce parameter was reviewed through responsible disclosure. Reproduction details remain private unless the owner authorises publication.
An Oracle-backed payment endpoint on a government-adjacent portal responded to a timing payload with a measurable delay, confirming injection. Disclosed privately and verified fixed. The target stays anonymous here.
Assessment evidence covered common web-security and configuration risks, with a prioritised fix roadmap. Detailed findings remain owner-review material.
A control-by-control gap review alongside live testing that surfaced stored XSS on search and review pages, an XML-RPC brute-force path and JavaScript libraries carrying known CVEs.
Scope reading and attack-surface mapping on inDrive, plus active reconnaissance on the Chime and TheFork programs — always inside published program rules.
A full OWASP Top 10 (2021) assessment across the client's web estate, reported with severity ratings and developer-ready remediation notes.
I started in commerce, which taught me early that a security failure is really a business failure. Then came a BS in Software Engineering, a stint running a city transit control centre, and in 2022 SCOLTECH — my own practice for security audits, Python automation and AI integration.
Today I'm doing an MSc in Computer Science at Xinjiang University, China, researching industrial AI where the factories actually are — and bringing that research discipline into every client report I write.
I write the finding, the fix, and the automation that keeps it fixed.
From China I cover Pakistan, Gulf and EU time, with US overlap.
Written for developers and executives in the same document.
Authorised, defensive, non-destructive testing. Always.
MSc research at Xinjiang University — making industrial and AI systems more precise, explainable and secure.
Visual anomaly detection for manufacturing: spotting defects on a production line and explaining them in a way a human operator can trust and act on.
Securing LLM pipelines and AI-powered products against prompt injection, data leakage, bias and supply-chain risk — the failure modes classic pentests miss.
Researching industrial AI and anomaly detection while running client security work.
Security audits, Python automation and AI integration for startups and SaaS teams. Built the VPN platform, KYAZZ ERP and 15+ tools.
Real-time monitoring and coordination of a city bus-rapid-transit network.
Best Employee of the Year 2022Earned the CNSS certification in 2020 and started bug-bounty hunting on HackerOne and Bugcrowd.
VERIFIEDVERIFIED20232023Public records and practice signals are kept separate from client claims. Certificate IDs and supporting details are available on request.
Yes, worldwide. My hours overlap Pakistan, the Gulf, the EU and part of the US day, and everything is delivered in English.
A free scoping call, then a written plan with fixed scope, timeline and price. Testing follows the OWASP Testing Guide and PTES. You get a CVSS-scored report with reproduction steps and remediation guidance, plus one free retest.
Yes — remote full-time, contract work and research partnerships, around 30+ hours a week. I can start immediately, and my CV is a message away.
Under NDA, inside the agreed scope only, with no client data retained after the engagement and reports shared encrypted.
Yes. References from kyazz.com, thebeehost and priceoye.pk are available on request, along with redacted report samples and live demos of my systems.
Bring one system or process to a free 20-minute call and I'll tell you honestly whether I can help — and what it would cost.
Usually replies within 24 hours · NDA on request · English