Turning raw data
into reasoning systems.
I'm Yassine, a third-year Data Science Engineering student at ESSAI. I build AI and Generative AI systems: LLM-powered products, RAG, and knowledge graphs. I like taking a messy real-world problem all the way to something that works.
From Tunis, building for a global AI industry.
I'm currently in my third year of an Engineering Degree in Data Science at ESSAI (École Supérieure de la Statistique et de l'Analyse de l'Information), after a Mathematics & Physics preparatory cycle at IPEIT.
My background started in classic data science and big data (dashboards, pipelines, statistics) and has moved deliberately toward Generative AI and LLM engineering: agents, retrieval-augmented generation, and knowledge graphs applied to real business problems.
I'm looking for an AI Engineering / Generative AI internship in France, Amsterdam, or elsewhere in Europe.
Engineering Degree in Data Science
Preparatory Cycle: Mathematics & Physics
Where I've applied this.
Tunisia
- Contributed to LedgerMind, a GenAI-powered platform simplifying financial and administrative tasks for content creators, enabling natural-language interaction with documents and financial data.
- Built intelligent workflows for tax estimation, financial simulation, invoice generation, and automated reporting from invoices and bank transfers.
- Applied Generative AI concepts (LLM applications, intelligent workflows, AI-assisted decision support) to a real business problem within a team.
Tunisia
- Worked within the Big Data team on data processing and analytics pipelines.
- Developed interactive statistical dashboards using Python and Power BI to support data-driven decision making.
Things I've built.
My flagship project, plus a range of applied ML and GenAI work.
🇹🇳 Tunisian Companies Intelligence: RAG → GraphRAG
A portfolio-grade AI system that answers questions over Tunisian companies' annual reports, starting as a baseline RAG pipeline and evolving into GraphRAG to reason across companies, sectors, strategies, risks, technologies, investments, markets, and time periods.
LedgerMind: GenAI Financial Assistant
AI-powered assistant helping content creators manage financial and administrative tasks through natural-language interaction with documents and data.
AIris: Real vs AI-Generated Image Detector
EfficientNetB0 with 3-phase transfer learning and Grad-CAM explainability, trained on CIFAKE to tackle deepfakes and synthetic media.
YouTube Tool-Calling Agent
An LLM agent built with LangChain that autonomously selects and invokes YouTube API functions to answer natural-language queries.
Baseball Data Analysis Agent
An AI analyst built with LangGraph and Pandas that answers natural-language questions over World Series statistics.
Customer Review Sentiment Analyzer
Sentiment classification (positive/negative/neutral) on Amazon and Yelp reviews via TF-IDF, word embeddings, and a transformer model.
Customer Churn Predictor
Binary classification on telecom data (contract type, charges, tenure) to predict churn before it happens.
Image Segmentation: Mean Shift Clustering
Color-based image segmentation using Mean Shift Clustering, completed as part of the IBM ML Professional Certificate.
Exploratory Data Analysis for Data Science
Comprehensive EDA on the Diabetes (regression) and Iris (classification) datasets.
What I work with.
Programming
AI / Machine Learning
Generative AI
AI Engineering
Backend / Development
Databases
Cloud / DevOps
Data & Analytics
Currently developing skills in: Advanced RAG · GraphRAG · LLM Evaluation · MCP · LLMOps