Armend Ujkani

Professional Biography & Technical Background

Armend Ujkani is a professional web developer, systems architect, and digital creator based in Prishtina, Kosovo. With a versatile technical foundation spanning PHP, WordPress architecture, modern HTML5, CSS3, ES6+ JavaScript, and Python, Armend has spent more than fifteen years developing robust web platforms, media management engines, and responsive digital solutions for international and domestic clients.

In his primary role as a PHP Developer at PixelCode Studio, Armend specializes in engineering bespoke, high-performance, and client-tailored web applications. His work emphasizes mobile-first responsiveness, clean procedural and object-oriented architectures, security compliance, and optimal Core Web Vitals.

Why I Built Clear AI Label: Defending Photographers Against Algorithmic False Flags

In early 2024, social media ingestion systems on Instagram, Facebook, and TikTok began deploying automated binary container scanners to identify synthetic media. These systems do not examine the visual pixels of a photograph; instead, they inspect JPEG header marker segments (specifically APP11 containing ISO/IEC 19566-5 C2PA JUMBF manifests) and Adobe XMP metadata (specifically iptcExt:DigitalSourceType).

As a direct consequence, genuine commercial and wedding photographers in my professional network started receiving false-positive "Made with AI" badges simply because they used minor retouching features—such as Adobe Lightroom's AI Denoise on a high-ISO wedding photo, or Photoshop's Generative Remove to eliminate a background electrical wire. These automated labels severely damaged creator reputation, client trust, and reach on social feeds.

Existing online metadata tools were either bloated, filled with deceptive ads, or required users to upload their high-resolution confidential photographs to third-party remote servers. I set out to engineer an elegant, honest, and technically rigorous alternative: Clear AI Label (clearailabel.com).

The Engineering Philosophy: 100% Client-Side, Zero Uploads

Clear AI Label was built under three uncompromising engineering rules:

1. Zero Remote Server Uploads: Every byte of inspection, binary parsing, and re-writing is executed locally in the visitor's web browser using JavaScript Uint8Array, DataView, and WebAssembly. No photo, PDF, or video is ever uploaded to our servers. This ensures complete GDPR and CCPA privacy compliance and commercial confidentiality.
2. Lossless Byte-Level Excision: For JPEG and PNG files, the cleaner operates directly on binary markers (e.g., excising APP11 marker segments 0xFFEB and open-weights PNG prompt chunks tEXt/iTXt) without recompressing pixel data, maintaining 100% of the original photo's visual fidelity.
3. Authentic Camera EXIF Preservation: In our signature AI-Only Mode, the engine parses the original Little-Endian binary TIFF structure (Make, Model, Lens, ISO, Focal Length, Exposure Time), removes only the AI provenance markers, and synthesizes a standards-compliant binary TIFF header so the file remains recognized as an authentic camera photograph.

Engineering Milestones & Changelog

Connect & Collaborate

I actively welcome feedback, feature suggestions, and container sample files from photographers, developers, and provenance researchers worldwide. You can connect with me directly through:

• LinkedIn: linkedin.com/in/armendujkani
• Technical Support & Bug Reports: [email protected]
• Digital Broadcasting: webradioguide.com
• YouTube Channel: youtube.com/@RrekiShqip