I’m Tesnime Ellabou, an AI Engineer, Data Scientist, and Researcher passionate about building intelligent systems that are robust, explainable, and impactful. I specialize in Machine Learning, NLP, Generative AI, Explainable AI (XAI), Deep Learning, and Big Data, blending technical precision with creative problem-solving. My work spans Arabic dialect TTS/STT systems, multimodal AI models, predictive analytics, and research in adversarial robustness and finance AI.
Machine Learning • Deep Learning • NLP & Transformers • Generative AI • Explainable AI (XAI) • Computer Vision • Big Data • MLOps • Predictive Analytics
- Investigated how subtle pixel-level perturbations can completely fool AI models, revealing vulnerabilities in state-of-the-art deep vision networks.
- Demonstrated the need for robust and trustworthy AI before real-world deployment.
- Highlighted ResNet18 accuracy collapse under FGSM & PGD attacks, emphasizing adversarial resilience.
- Developed a multi-domain model combining tabular financial indicators with NLP-driven signals to detect hidden anomalies.
- Integrated Explainable AI (SHAP) for transparent, accountable predictions in finance.
- Implemented production-ready MLOps workflow using MLflow & Weights & Biases for reproducibility and monitoring.
- 🧩 Building custom Text-to-Speech models for Tunisian Arabic
- 🧠 Fine-tuning LLMs with LangChain & domain-specific datasets
- ⚙️ Developing robust AI pipelines with MLOps best practices
- 📊 Scaling models for Big Data architectures and real-world deployment
Always happy to connect, collaborate, and explore AI innovations 🤝
“Elegant AI is not complexity , it’s clarity, robustness, and explainability.”