Younès BENSAFIA

MCP server for Overleaf projects. Syncs LaTeX files via Git, parses sections, equations, and citations, and exposes them to AI clients for assisted paper review, LaTeX fixes, and content generation.

Lost in a new codebase? DevLens scans your Python project and scores every file from 0 to 100 based on how hard it is to understand. It uses code complexity metrics, git history signals, and optional AI to pinpoint the files your team will struggle with. It also shows project statistics (languages, lines, directories) and summarizes files with AI.

Want to search arXiv papers, fetch metadata, and extract full-text PDFs without leaving your editor? This MCP server connects any MCP-compatible client (Claude Code, etc.) directly to arXiv.

A local server that lets AI assistants like Claude search, browse, and read full papers from arXiv, the giant open-access library of science and computer science research, without leaving your editor or chat window.

Tired of manually searching for remote developer jobs? This Python Discord bot scrapper collects and structures listings from Emploitic, making them ready for analysis or integration into dashboards and job boards

Block-matching motion estimation inspired by the one used in Twitch, YouTube, and Zoom. Feed two images, inspect motion vectors and residual heatmaps. Three-step search, batch processing, YCrCb/BGR/Y-only modes. Interactive PyQt6 GUI and CLI.

An immersive 3D museum experience developed using OpenGL 3.3. This project explores real-time rendering, dynamic lighting systems, and interactive gameplay mechanics within a high-fidelity virtual environment.

A collection of Game Theory concepts, models, and implementations exploring strategic decision-making, equilibria, and rational behavior in competitive environments.

This repository is demonstrating network programming concepts using Java sockets, covering both TCP and UDP protocols, as well as packet manipulation.

This repository contains resources and examples used in my workshop for Algeria 2.0 for creating your own mini AI workshop using large language models (LLMs) and datasets from Hugging Face.

Can the knight visit every square on the chessboard exactly once?, This is where the genetic algorithm steps in: by evolving generations of possible tours, selecting the most promising ones, it gradually guides the knight toward the perfect journey one that covers every square exactly once

A collection of interactive data visualizations built with D3.js, showcasing various chart types, statistical methods, and analytical techniques for visual data exploration.

Model Context Protocol Servers

A workshop for GDG Algiers as part of the GDG Integration Program, where I explained the essentials and fundamentals of Git and GitHub.

A compiler for the PHYLOG language, built with FLEX and BISON, featuring lexical, syntax, and semantic analysis, symbol table management, error handling, and object file generation.

This repository focuses on facial landmark extraction, preprocessing, and analysis using PCA (Principal Component Analysis / ACP). It provides tools to process a facial dataset, extract meaningful geometric and statistical features, and perform dimensionality reduction for visualization and further analysis.

This project focuses on developing a traffic sign recognition system using the Scale-Invariant Feature Transform (SIFT) algorithm. The goal is to detect and recognize traffic signs in images with diverse backgrounds

The New York City Airbnb Data Cleaning Project aims to preprocess and clean the publicly available Airbnb dataset for New York City to make it reliable and ready for analysis or machine learning tasks.

How would the Minimax algorithm choose its very first move in Mancala? Mancala, also known as Awalé, is an African-origin sowing game played on a board with six small pits on each side and a store at each end. Each pit starts with four seeds, and the challenge is to sow them strategically to outscore your opponent

A Web3 development workshop project that demonstrates how to build decentralized applications (DApps) using Motoko on the Internet Computer. It serves as a practical introduction to smart contracts, state management, and Web3 concepts.

A comprehensive collection of Java multithreading and concurrency examples demonstrating fundamental thread synchronization patterns, the producer-consumer problem, and advanced multi-queue systems.

Code at the speed of thought – Zed is a high-performance, multiplayer code editor from the creators of Atom and Tree-sitter.

Display and control your Android device

Documentation for GDG's code guidelines and workflow

Enhanced Apriori rule mining with a custom asymmetric quality function using tunable α and β to control penalties for false positives and false negatives. Includes comparison with traditional metrics like confidence and lift.

Implement three algorithms to check if a number n is prime. Test each algorithm with different values of n, measure performance, and compare results visually.

An ongoing, collaborative meta-analysis about Human-AI-Interactions. We aggregate data and knowledge to build a non-abrasive, user-friendly prompting framework tailored to LLM mechanics, ensuring reasoning stability and a friction-free prompting environment that is safe for the human psyche and wellbeing.

CS50x-Problem-Sets: Solutions to Harvard's CS50x course problem sets, exploring core computer science concepts like algorithms, data structures, C, Python, SQL, and web development.

DoIt! is a web application designed for managing tasks, likely inspired by common "To-Do List" apps. It helps users organize and keep track of their daily or long-term tasks. This project is the final submission for the Harvard CS50x course.

This repository is dedicated to my learning journey in linear regression. I'm following tutorials, experimenting with different datasets, and implementing regression models to deepen my understanding.

This repo is my personal collection of books, articles, and resources about Machine Learning and Deep Learning. Whether you're just getting started or diving deep into advanced topics, you'll find something valuable here.

Parallel-Quick-Sorting is a repository showcasing the implementation of the Quick Sort algorithm optimized for parallel processing.

"Complexity-Practices" is a repository dedicated to exploring algorithms, and computational complexity through hands-on exercises and practical implementations.

This repository explores the construction of a suffix array using various sorting algorithms (Quick Sort, Heap Sort, Merge Sort, Bucket Sort) and analyzes their practical complexities. It also includes a study of the search complexities within the suffix array.

No description.

This repository contains Python implementations for random data generation, search algorithms, sorting, matrix generation, matrix multiplication, and matrix power computation, along with complexity analysis.

This repository contains two different algorithms designed to compute a specified sum. The purpose is to compare their time complexities through practical measurements by varying the size of n.

Assembly project demonstrating interrupt-driven task scheduling: single-task mode with 1-second timer and multi-task mode with 5 tasks triggered every 5 seconds, featuring modular interrupt management.

Fine-tuning & Reinforcement Learning for LLMs. 🦥 Train OpenAI gpt-oss, DeepSeek-R1, Qwen3, Gemma 3, TTS 2x faster with 70% less VRAM.

A lightweight, powerful framework for multi-agent workflows

Examples and guides for using the OpenAI API

in this repo we are resolving SokobanPuzzle using algorithms .

Ultimate Solidity, Blockchain, and Smart Contract - Beginner to Expert Full Course | Python Edition

An AI coding agent built in Python with the Gemini Flash API. Features an agentic loop with tool calling for reading, writing, and executing code to iteratively improve a codebase.

Python tool for converting files and office documents to Markdown.