Md Shameem Alam.

Developed and deployed an online judge system with React.js frontend, Node.js/Express backend, Redis queue, and Docker compiler on AWS EC2/ECR, reducing execution latency by 15%. Implemented real-time verdicts with SSE, JWT authentication, and AI-powered code review for load tested up to 10 concurrent users using k6. Strengthened execution security and scalability by implementing sandboxed environments, enabling safe, concurrent compilation across 5+ programming languages and reducing runtime failures by 20%.

Engineered a hybrid recommender system by integrating popularity-based, collaborative, and content-based models, enabling personalized book suggestions for 10k+ users. Streamlined preprocessing of 100k+ records, correcting missing values, improving reliability and integrity by 20%. Optimized deployment pipeline by storing trained models and similarity matrices as serialized artifacts, improving scalability, reducing model load time by 40%, and cutting query time by 35%. Achieved 25% improvement in recommendation accuracy, mitigating cold-start issues and enhancing personalization.