I’m an Artificial Intelligence engineering student at ENIB (École Nationale d'Ingénieurs de Brest), passionate about speech synthesis, signal processing, computer vision, and building offline intelligent systems.
In my free time, I design and develop fully local voice assistants, machine learning applications for voice, and innovative projects at the intersection of cybersecurity, computer vision, and AI-generated content detection (like deepfake audio).
I specialize in creating privacy-focused local voice solutions and explore open-source LLMs like GPT4All, Nous Hermes, and Mistral. I also enjoy working with tools like FastAPI, Streamlit, and OpenCV to build real-time AI-powered interfaces.
- 💻 AI & Machine Learning: Real-time models, voice AI, local-first systems
- 🎤 Speech Synthesis: Custom voice assistants, voice-to-voice pipelines
- 👁️ Computer Vision: Fatigue detection, facial analysis, eye tracking
- 🛡️ Cybersecurity: Deepfake detection, audio anti-spoofing
- Speech Synthesis: Tacotron2, VITS, Coqui TTS, SoX
- Speech Processing: Emotion detection, pitch/tempo modulation
- LLMs: GPT4All, Nous Hermes, Mistral
- Computer Vision: OpenCV, Dlib, Mediapipe
- Deep Learning: PyTorch, TensorFlow, CNN, RNN
- AI Frameworks: FastAPI, Streamlit, Gradio, Hugging Face
- Languages: Python, C, Java, SQL, Bash,
- Web/API: Django, React, HTML5, CSS, PostgreSQL
- Tools: Git, GitHub, GitLab, Jupyter
- Deepfake audio detection
- Local-first AI deployments
- French (Native)
- English (Fluent)
-
Numa
A fully local intelligent voice assistant combining speech recognition, open-source LLM, and expressive speech synthesis — all without cloud dependencies. -
FatiguEye
A fatigue detection app using computer vision to track eye blinks and micro-sleep indicators via webcam (EAR-based analysis). -
EmotionAI-Voice
A tool that detects emotions (stress, calm, anger...) from voice recordings using deep learning and a user-friendly interface.
