Franck Ngaha
Senior Python & AI Engineer
Building production-ready AI systems for real-world business decisions.
Value I bring
- Production mindset: deployable APIs, CI/CD, Docker, Cloud Run
- Business-first ML: decisions, thresholds, costs, review zones
- Explainability & auditability for regulated environments
What I do
Production-ready AI & ML systems
From data pipelines to deployed APIs, I build machine learning systems designed to run in production — not just notebooks.
Business decision layers on top of ML
I design explicit decision rules (accept / review / reject, thresholds, business constraints) on top of probabilistic models.
The model predicts risk. The business keeps the decision.
Explainable AI for regulated environments
Interpretable models and explainability techniques to make decisions understandable, auditable, and defensible.
Python engineering & data pipelines
Clean Python foundations for APIs, data processing, and ML workflows — structured, testable, and production-oriented.
Featured project
Credit Risk Scoring — Business Decision Thresholds
From ML probability to actionable credit decisions
An end-to-end ML system where a model estimates P(bad) and a business layer converts it into ACCEPT / REVIEW / REJECT using configurable thresholds.
- Dataset: OpenML credit-g
- Model: Logistic Regression (interpretable & stable)
- ROC AUC ≈ 0.78 (as support signal, not the end goal)
- Decision thresholds optimized with weighted business cost
- Deployed on Google Cloud Run
How I work
I build systems that are easy to operate, explain, and evolve — not prototypes that stop at notebooks.
My focus is to turn model outputs into clear decision logic, with guardrails, monitoring, and production constraints in mind.