👋 Hi, I'm Muneeb

Engineered a dual-model NLP pipeline for confidence based Disaster Tweet classification. Built a production-oriented MLOps cycle covering model inference, API deployment, data persistence, monitoring, and prediction tracking using FastAPI and Firebase Data Connect (PostgreSQL).

Built a content + collaborative-filtering movie discovery system using FAISS and LightFM.

Integrated a Deep Q-Network (DQN) agent into a C++ game using LibTorch, successfully stabilizing a 55-dimensional state space and 15-action space using experience replay, epsilon-greedy exploration, and target network updates.

Built a deep learning classifier from in C++, implementing neural network layers, backpropagation, and model serialization, achieving 85%+ accuracy on MNIST.

Comprehensive survey analysis of 100 students examining the relationship between social media usage and mental well-being. Key finding: 51% reported mental health decline after increased usage.

Secured 7th position in CSP solver competition for logic grid puzzles (Zebra puzzles) using symbolic reasoning, not ML. Implemented backtracking search with MRV heuristic, degree heuristic as tie-breaker, Least Constraining Value ordering, and AC-3 arc consistency for early dead-end detection.

Healthcare management system built with Java, Spring Boot & JavaFX for real-time patient vital monitoring, automated medical report generation, appointment scheduling, and emergency alerting through email.

Quantum-resistant messaging app using ML-KEM (Kyber) post-quantum cryptography for secure communications over a local area network (LAN).

Google Docs-inspired web editor to create and export documents as .docx files.

Built a desktop job portal connecting employers with tradespersons and laborers through city and skill-based job matching, backed by CustomTkinter and Firebase Firestore.