Karim Badreldin@Kimo
1
Events attended
1
Submissions made
Germany
2+ years of experience
About me
I have a Master’s degree in Software Engineering from the University of Stuttgart, specializing in Artificial Intelligence, I bring a strong foundation in both advanced AI theory and practical software development. My academic and professional journey has been centered on AI-driven systems, including deep learning, computer vision, and reinforcement learning. I’ve developed cutting-edge models for multi-human pose forecasting using temporal and attention-based architectures, as well as real-time object detection, tracking, and path planning for autonomous vehicles. Professionally, I’ve delivered impact across roles at TeamViewer and Gameball—contributing to secure system architecture, CI/CD pipeline automation, and API integration. My experience spans both enterprise- scale and startup environments, reflecting adaptability and a results-oriented mindset. With strong analytical skills, deep knowledge of AI and machine learning frameworks, and proficiency in scalable software engineering, I am dedicated to building intelligent, secure, and high-performing systems.
🤓 Latest Submissions
Private Virtual Stylist
Private Virtual Stylist is a smart, on-device fashion assistant that allows users to virtually try on clothes and receive intelligent outfit recommendations—all while ensuring total privacy. Built with the HR-VITON pipeline for high-quality try-on synthesis, the system uses Grok to extract metadata from clothing images (e.g., type, color, pattern) and LLaMA to generate natural language fashion advice tailored to the user's current outfit or selected item. Running entirely offline on Qualcomm Snapdragon-powered Edge AI devices, this solution is optimized for privacy-first fashion experiences—ideal for mobile shopping, in-store kiosks, or use in regions with limited connectivity. Key Capabilities: • Virtual Try-On: Upload a photo and clothing item → see realistic try-on result • Style Intelligence: Metadata extraction via Grok + outfit advice from LLaMA • Offline-First: All inference runs on-device—no cloud, no data sharing • Privacy-First: Aligned with GDPR and secure by design Tech Stack Highlights: • HR-VITON for image-based try-on • Grok for clothing metadata parsing • LLaMA (LLM) for outfit suggestions and personalized recommendations • Pose estimation & image preprocessing for realism • Edge AI deployment using ONNX
8 Jul 2025
