Tajeddine Bourhim - Data Scientist & AI Engineer

Featured Projects

A showcase of data science and AI projects that demonstrate expertise in machine learning, big data processing, and intelligent system development.

Full-Stack AI Development

MCP Project Spring Python

A comprehensive full-stack implementation of the Model Context Protocol (MCP) featuring a Spring Boot backend client, Python MCP server, and Angular frontend. Enables seamless AI-powered chat interactions with integrated tools for real-time stock market data, company financials, file system operations, and employee database management through a unified conversational interface.

Purpose: Develop a sophisticated MCP ecosystem demonstrating advanced AI tool integration, multi-language architecture, and real-time data processing capabilities to showcase enterprise-grade AI application development.

Approach: Architected a microservices-based solution with Spring Boot handling RESTful API endpoints and business logic, Python server managing MCP protocol communication and AI tool orchestration, and Angular providing a modern, responsive user interface. Integrated multiple AI tools including financial data APIs, file system utilities, and database operations with proper error handling and real-time data synchronization.

Jun 2025 – Jul 2025

Full-Stack Developer & AI Integration Specialist

Outcome: Successfully delivered a production-ready MCP platform with seamless AI tool integration across multiple programming languages, demonstrating expertise in distributed systems, API design, and modern web development practices.

Spring BootPythonAngularMCPAI tool integrationOllamaREST APIsMicroservices

Code

AI-Powered Medical Diagnostics

MediScan - AI-Powered Pneumonia Detection System

A comprehensive medical diagnostic platform leveraging state-of-the-art deep learning models to analyze chest X-rays with 94.2% accuracy and 93.8% sensitivity. Features include Grad-CAM visualization for interpretability, AI-generated clinical reports, patient management system, batch processing capabilities, and HIPAA-compliant secure data handling. Built with Next.js frontend, Flask backend, TensorFlow AI models, Docker containerization, and MongoDB database.

Purpose: Develop an advanced AI-powered diagnostic tool for rapid and accurate pneumonia detection from chest X-rays to assist healthcare professionals in early diagnosis, improve patient outcomes, and streamline clinical workflows in hospitals and medical facilities.

Approach: Implemented using Next.js 14 with App Router for responsive frontend, Flask backend with TensorFlow EfficientNet-B0 model for AI processing, MongoDB for secure patient data storage, and Docker for containerized deployment. Integrated Grad-CAM for model interpretability, JWT authentication for security, and comprehensive API endpoints for medical workflow integration. Includes automated Docker management scripts for cross-platform deployment and detailed clinical documentation generation.

Mar 2025 – Jun 2025

AI/ML Engineer & Full-Stack Developer

Outcome: Successfully developed a production-ready medical diagnostic platform achieving 94.2% accuracy with 3-5 second processing times. Features include interactive Grad-CAM heatmaps, comprehensive clinical report generation, multi-patient dashboard, batch processing for high-volume workflows, and HIPAA-compliant data security. Deployed with Docker containerization supporting cross-platform installation and includes detailed API documentation for healthcare system integration.

Next.jsFlaskTensorFlowDockerMongoDBEfficientNet-B0Grad-CAMJWTTailwind CSSshadcn/ui

Code

AI & NLP

Chat with LLaMA-2 and PDFs

An intelligent conversational AI system that enables natural language interaction with PDF documents through advanced semantic search and question-answering capabilities. Leverages state-of-the-art language models to provide contextually relevant answers, document summarization, and real-time insights extraction from uploaded PDF files.

Purpose: Create an advanced document intelligence platform that transforms static PDF content into interactive, queryable knowledge bases using cutting-edge natural language processing and retrieval-augmented generation techniques.

Approach: Developed a robust pipeline using Python and Streamlit for the user interface, implementing FAISS vector database for efficient semantic similarity search, Hugging Face Transformers for LLaMA-2 model integration, and PyPDF2 for comprehensive document parsing. Incorporated advanced text preprocessing, chunking strategies, and embedding techniques to optimize retrieval accuracy and response quality.

Dec 2024 – Jan 2025

AI/ML Engineer & NLP Specialist

Outcome: Delivered a fully functional AI-powered document chat system capable of processing complex PDF documents and providing accurate, context-aware responses, demonstrating expertise in large language models, vector databases, and information retrieval systems.

PythonStreamlitFAISSHugging Face TransformersLLaMA-2PyPDF2Vector SearchNLP

Code

Data Science & Finance

Cryptocurrency Price Prediction

A comprehensive cryptocurrency analytics and forecasting platform featuring real-time market data visualization, advanced price prediction models, and interactive trend analysis tools. Provides traders and investors with data-driven insights through machine learning algorithms and comprehensive market indicators.

Purpose: Develop a sophisticated financial analytics platform that combines real-time market data with predictive modeling to help cryptocurrency traders make informed investment decisions through data-driven insights and automated forecasting.

Approach: Built a full-stack application using Python and Streamlit for the interactive dashboard, integrating CoinGecko API for live cryptocurrency data feeds, implementing multiple Scikit-learn regression models (Linear Regression, Random Forest, LSTM) for price prediction, and utilizing Altair for advanced data visualizations. Incorporated technical indicators, market sentiment analysis, and automated model retraining capabilities.

Nov 2024 – Dec 2024

Data Scientist & Financial Technology Developer

Outcome: Successfully launched a production-ready cryptocurrency analysis platform with real-time data integration, multiple prediction algorithms, and comprehensive visualization tools, demonstrating expertise in financial data analysis, time series forecasting, and fintech application development.

PythonStreamlitCoinGecko APIScikit-learnLSTMRandom ForestAltairTime Series Analysis

Code

Computer Vision

Face Recognition System

A high-performance real-time facial recognition and attendance tracking system achieving 90% accuracy with sub-2-second processing times. Features automated face detection, recognition, and attendance logging capabilities suitable for educational institutions, corporate environments, and security applications.

Purpose: Develop a robust computer vision solution for automated attendance management and access control through accurate facial recognition technology, addressing the limitations of traditional attendance methods while ensuring privacy and security.

Approach: Implemented a comprehensive computer vision pipeline using Python and OpenCV, featuring Haar cascade classifiers for efficient face detection, K-Nearest Neighbors algorithm for facial recognition with optimized feature extraction, and SQLite database for secure attendance record management. Incorporated real-time video processing, face alignment techniques, and anti-spoofing measures for enhanced security and reliability.

Apr 2024 – Jun 2024

Computer Vision Engineer & System Architect

Outcome: Successfully deployed a production-ready face recognition system with 90% accuracy and real-time performance, demonstrating expertise in computer vision algorithms, image processing techniques, and practical AI system implementation for enterprise applications.

PythonOpenCVHaar CascadeKNN AlgorithmSQLiteReal-time ProcessingComputer Vision

Code

Web Scraping & Data Aggregation

Moroccan News Aggregator

A comprehensive web scraping and data aggregation platform that collects, processes, and categorizes news articles from multiple Moroccan news sources. Features automated content extraction, intelligent categorization, and cloud-based storage with an intuitive interface for accessing localized news content.

Purpose: Create a centralized information hub for Moroccan news consumers by developing an automated web scraping solution that aggregates content from diverse sources, enabling better access to local news and supporting data journalism initiatives.

Approach: Developed a robust data pipeline using Streamlit for the user interface, implementing Selenium WebDriver for dynamic web scraping across multiple Moroccan news websites, integrating Google Drive API for scalable cloud storage, and utilizing Pandas for data processing and cleaning. Incorporated intelligent content categorization, duplicate detection algorithms, and scheduled automated updates to ensure fresh and relevant news delivery.

July 2024 – Sep 2024

Data Engineer & Web Scraping Specialist

Outcome: Successfully built and deployed a functional news aggregation platform serving multiple Moroccan news sources with automated scraping capabilities, cloud storage integration, and user-friendly data access, demonstrating expertise in web scraping, data engineering, and automated data pipeline development.

StreamlitSeleniumGoogle Drive APIPandasPythonWeb ScrapingData ProcessingCloud Storage

Code

View All Projects on GitHub