Usman Nazeer

AI/ML Developer and Full-Stack Engineer with experience building FastAPI backends, computer vision systems, and AI-integrated web applications. Comfortable working across the stack, from database design and API development to model integration and deployment.

AI Full Stack Intern at Nexium (2025-06 – 2025-08)

• Built full-stack web applications with React.js, Node.js, and MongoDB, integrating AI-powered features with the Gemini API and n8n to enhance user engagement.
• Collaborated with UI/UX teams on modern design improvements and automated deployments using Git and CI/CD tools like Vercel, reducing deployment time and errors.

AI/ML Developer at SparX Technologies (2025-08 – Present)

• Developed a FastAPI claims management system with MySQL connection pooling, REST endpoints, and multi-company scheduling, boosting processing speed and reliability.
• Built a geospatial inspector-matching algorithm using haversine distance, LRU caching, and weighted scoring, achieving a 60% latency reduction, alongside an ETL pipeline processing 1,000+ CSV records with fuzzy mapping and Azure integration for 98% data integrity.
• Implemented image fraud detection with perceptual hashing, OpenCV matching, EXIF forensics, and blur analysis, achieving 95% authenticity confidence.
• Built a FastAPI backend for AI-driven property walkthrough documentation, featuring automated report generation with GPT-4o image analysis and compressed PDF export, using PostgreSQL and Redis for session and report management, which accelerated report delivery and reduced manual effort.
• Built an end-to-end roof analysis platform with a FastAPI backend and React frontend, integrating a trained YOLO-based segmentation model to process satellite imagery for automated roof detection and area estimation, delivering faster, consistent assessments for planning teams.

Bachelors of Science in Computer Science – National University of Computer & Emerging Sciences Islamabad, Pakistan (2021-08 – 2025-06)