Hello, innovators! I'm Wiem Zakraoui, a Data Science and AI enthusiast dedicated to building intelligent, safe, and impactful solutions. My GitHub is a hub for projects in Natural Language Processing (NLP), Large Language Models (LLMs), MLOps, and AI-driven automation, designed to solve real-world challenges. Ready to explore? Let’s code the future together!

• NewsBot IA – LLM-Powered News Analysis
  A Python-based AI assistant for journalists, leveraging LLMs to detect breaking news, summarize articles, and flag risk-related content (e.g., fraud, sanctions). Built with Hugging Face, BeautifulSoup, Scrapy, and FastAPI, deployed with Docker and ELK for real-time alerts. Implemented safety layers for prompt injection prevention and hallucination control.
  Why it matters: Delivers fast, trustworthy insights for high-stakes environments, aligning with compliance needs.

• RGAA VS Code Extension – LLM-Driven Accessibility Audits
  Developed a VS Code extension powered by an LLM to automate RGAA accessibility audits for web content, detecting 50+ anomalies with advanced prompt engineering. Integrated via FastAPI and deployed in a CI/CD pipeline using Docker, Jenkins, and Grafana in an Agile Scrum environment.
  Why it matters: Enhances digital accessibility with efficient, scalable AI automation.

• Automated MLOps Pipeline for Model Deployment
  An end-to-end MLOps pipeline using MLflow, Jenkins, and Docker to automate NLP model training, versioning, and deployment. Reduced deployment time by 80% with continuous monitoring via Prometheus.
  Why it matters: Streamlines AI workflows for reliable, production-ready solutions.

• Deep Learning for Brain MRI Classification
  A deep learning project using ResNet50 to classify brain tumors from MRI scans, achieving 98.2% accuracy. Built with PyTorch and integrated text-based reporting for medical insights.
  Why it matters: Advances healthcare with precise, data-driven diagnostics.

I’m a Data Science engineering student at ESPRIT, specializing in Artificial Intelligence. My expertise includes:

• Conversational AI & LLMs: Designed AI assistants with Hugging Face, SpaCy, and FastAPI, incorporating AI safety (data masking, hallucination control).
• MLOps & Deployment: Automated pipelines with Docker, Jenkins, and MLflow for scalable AI solutions.
• Technical Stack: Proficient in Python, TensorFlow, PyTorch, SQL, AWS, and GCP.
• Professional Experience: Interned at Talan Tunisia, developing an LLM-powered VS Code extension, and at DRÄXLMAIER Group, delivering actionable data insights. Thrived in Agile Scrum environments.

I’m passionate about creating responsible AI that empowers users, from financial services to healthcare, and I’m excited to tackle challenges in conversational AI and compliance.

Have an idea for an AI project or want to enhance conversational systems? Open an issue, star a repo, or reach out to collaborate. Let’s build something impactful!

• Email: wiem.zakraoui@esprit.tn
• LinkedIn: linkedin.com/in/wiem-zakraoui
• GitHub: github.com/wiemtn

Last updated: November 2025

⭐ Star a repo to support my work and let’s create tech magic together! ⭐