Muhammad Usama U

Designed, maintained, and extended X-LES, a large-scale C++ research framework for Large Eddy Simulation (LES) and turbulent fluid flow analysis. Refactored legacy code to improve modularity and maintainability, ensuring strict adherence to internal coding standards and documentation. Developed technical components for mesh generation, mathematical utilities, and file I/O using NetCDF for structured scientific data storage. Enhanced system reliability by authoring comprehensive unit tests in Catch2 and managing incremental updates in a GitLab environment. Collaborated with researchers to translate complex simulation requirements into verified, high-performance C++ implementations.

Built and scaled an AI-powered distributed NestJS (TypeScript) platform for cross-retailer product discovery, matching, and price tracking across 80+ e-commerce retailers, designed to handle anti-bot protections at scale. Designed and integrated a Python/Django scraping service alongside AI-assisted crawlers, creating a hybrid scraping architecture that reduced third-party scraping infrastructure costs by 32%. Built cross-platform product matching to identify identical products across retailers and map pricing and availability despite inconsistent titles, markup, and catalog structures. Leveraged AWS Lambda for on-demand crawling and deployed the platform in Docker on AWS EC2 behind an Application Load Balancer with Redis/SQS, PostgreSQL, S3, VPC/private subnets, and CloudWatch monitoring. Integrated partner-seller and affiliate monetization flows across retailers including Backcountry, SkiEssentials, and Utah Skis using Impact, AvantLink, and Shopify. Developed React-based dashboards

Built and delivered multiple client products in AI, healthcare, fintech, education, and civic tech, implementing production systems with Django, .NET, NestJS/Node.js, React, AWS, Azure, and Jenkins. Owned technical planning, estimation, solution architecture, CI/CD, and environment management across development, staging, and production for multiple concurrent client engagements. Built Mincy, an AI-powered avatar customer support experience system for barbershops using AWS AI services for facial recognition, speech-based workflows, preference tracking, and appointment-related interactions. Engineered platforms including Fractal Vision and Solace Health, delivering teacher/student workflows, appointment booking, patient history, and provider-facing experiences using modern full-stack web technologies. Built workflow and scoring systems for DRG Invest and University of Colorado Boulder, including a party registration platform contributing to a 25% reduction in party-related ticket issues.