- 🔭 Building production-grade Machine Learning & AI systems
- 🧠 Specialising in Deep Learning · Reinforcement Learning · Topological Data Analysis
- 📊 Focused on robust data pipelines, rigorous model evaluation, and applied AI
- 🚀 Next up: real-time ML apps, trading/forecasting models, and deployed web AI
| Project | Description | Stack |
|---|---|---|
| TDA Supply Chain | Real-time topological anomaly detection for global supply chains using persistent homology (H₀ H₁ H₂), Wasserstein distance scoring & CUSUM adaptive thresholding. FPR < 1%, detects port blockages, fraud rings & geopolitical shocks within 5 min. | Python · FastAPI · Docker · TDA |
| Actor-Critic + PPO Trainer | Policy optimisation and reward convergence experiments with modular RL training loops | PyTorch · RL |
| Glaucoma Detection DCNN | CNN-based retinal image classification for early glaucoma detection from fundus images | TensorFlow · Medical AI |
| Crypto Benchmarks | Performance analysis of crypto assets — regression, PCA & clustering on financial time-series | Pandas · Scikit-learn |
| Leave-One-Out CV | Clean implementation of LOOCV and core model evaluation techniques | Python · Scikit-learn |
Languages Python · C · C++
ML/DL TensorFlow · PyTorch · Scikit-learn · Keras
Data Pandas · NumPy · Jupyter · Matplotlib · Seaborn
MLOps FastAPI · Docker · GitHub Actions
Math Topological Data Analysis · Persistent Homology · Linear Algebra · etc
📧 chegenganga08@gmail.com · 🐙 github.com/Chege-N
Perpetually probing internals & shipping solutions that compress complexity into code.