Welcome to my GitHub. I'm currently a student at CentraleSupélec (top-tier French engineering school), pursuing my gap year with a focus on applied machine learning and quantitative research. My projects explore the intersection of artificial intelligence, brain-inspired computation, and financial systems.
- Applied Machine Learning: NLP, time series, speech, and structured data
- Neuromorphic Computing: Reservoirs, dynamics, and alternative computing models
- Quantitative Finance: Strategy simulation, market analysis, portfolio modeling
- Cognitive Systems: Learning mechanisms inspired by biological intelligence
| Category | Tools / Frameworks |
|---|---|
| Languages | Python (Advanced), C++ (Intermediate) |
| ML Frameworks | PyTorch, scikit-learn, Hugging Face Transformers |
| Tools | Git, Docker, VS Code, Jupyter, SQL |
| Libraries | NumPy, OpenCV, Matplotlib, Optuna, Librosa |
| Domains | NLP, Time Series Forecasting, Speech Processing, Neuromorphic Models, Financial Modeling |
A neuromorphic learning system simulating a quantized photonic reservoir, applied to MNIST classification and stability analysis. Includes Lyapunov-like divergence metrics, leaky integration models, and Bayesian hyperparameter optimization.
Result: 98.86% test accuracy on MNIST (HOG + 8000-neuron reservoir)
A blindfold chess interface powered by speech recognition. Includes natural language translation of moves, fallback to engine-based evaluation, and a dynamic graphical interface for post-game review.
Focus: Real-time NLP, state tracking, and human-in-the-loop decision systems
A modular framework for backtesting and simulating trading strategies. Includes classical strategies (momentum, breakout), market cycle detection via FFT, an SVM baseline, and a virtual trading simulator.
Extension: Reinforcement learning agent (PPO) with BTC/USD series (excluded from Git)
A robust PDF CV parsing pipeline combining a fine-tuned CamemBERT model for named entity recognition with a T5-based text normalization module. Designed for robustness across diverse document formats and multilingual input.
- How intelligent behavior emerges from simple rules (biological or artificial)
- Market structure, alpha generation, and investment strategy modeling
- High-efficiency AI systems for constrained environments
- Bridging theory and practical deployment in machine learning
I'm actively seeking opportunities that involve:
- Applied AI or ML research
- Quantitative modeling and strategy design
- Internships or collaborations at the frontier of AI and finance
Feel free to reach out if you're working on something meaningful involving AI, cognition, or finance.
- Email: [email protected]
- LinkedIn: paul-lemaire
- GitHub: @Paul92150