Kijumbe James

I'm a Data Scientist and AI/ML Engineer based in Dar es Salaam, Tanzania, with over 2 years of hands-on experience designing, training, and deploying deep learning models and full-stack systems end-to-end. I specialize in translating real-world community and business challenges — such as crop disease detection and clinic workflow automation — into production-ready, offline-first software. My standout achievement is training and benchmarking EfficientNetLite0 and MobileNetV2 image classifiers from scratch on a 61,184-image, 19-class cassava and maize disease dataset, reaching 91.5% accuracy and 0.93 macro F1. I then optimized the winning model into a 2.88MB INT8 TFLite build with ~11.4ms inference, enabling real-time offline diagnosis on low-end Android devices through LeafScan TZ, an app built for Tanzanian smallholder farmers. On the full-stack side, I've deployed 10+ projects using Flask, Node.js/Express, MongoDB, MySQL, and PostgreSQL across Vercel and Render, including healthcare management systems, inventory platforms, and AI chatbot interfaces built with Next.js and React. I also bring strong data analysis and BI capabilities with Power BI and Advanced Excel. I'm now seeking remote roles in AI/ML engineering, full-stack development, and data science — open to freelance, contract, and full-time opportunities where I can build impactful, scalable technology from anywhere.