Omar Ramadan Ramadan

Led a cross-functional team as Team Leader to design and build SENTRA, a multi-modal AI-powered interview coaching platform that evaluates candidates through Computer Vision, Speech Processing, and Large Language Models (LLMs). Owned the Analysis & Business Model track: defined the platform's value proposition, candidate-evaluation criteria, and end-to-end user workflow for the coaching experience. Built the Computer Vision module using YOLOv8 for real-time facial and emotion analysis during mock interview sessions. Implemented Speech Processing using Whisper for real-time, accurate transcription of candidate responses. Contributed to platform-wide AI integration: LLaMA3.3 for adaptive interview question generation, ChromaDB with a knowledge base of 250,000+ interview Q&A pairs, Retrieval-Augmented Generation (RAG) for context-aware feedback, Prompt Engineering for personalization, and FastAPI for backend integration.

Built an end-to-end spam detection pipeline with text preprocessing, TF-IDF vectorization, and trained Naive Bayes and Logistic Regression classifiers. Evaluated models using accuracy, precision, recall, and confusion matrix for comprehensive performance comparison.