Quantmetry (Capgemini Invent) · 2019 · Data & Solution Architect
MLOps (re)training automation — white paper & talk
Authored chapter 3 of Quantmetry's 5th white paper, "AI in production: life cycle & model drift" — "Scheduled updates & (re)training management" — covering MLOps (re)training automation across code, architecture and delivery, with a 20-minute companion talk.
- Published reference on operationalizing ML — model life cycle, drift detection and automated retraining.
- Framed the architecture and delivery patterns that turn a model into a maintainable production system.
- MLOps
- Model drift
- CI/CD
- Cloud architecture
The chapter below — "Évolutions programmées et gestion du retrain" (in French) — covers how to keep a model reliable once in production: scheduling updates, detecting model drift, and automating (re)training across code, architecture and delivery. Read the full white paper (PDF, 14 MB).
The companion talk: a 20-minute walkthrough of the chapter — model life cycle in production, drift detection and automated (re)training — and the patterns that make it work in practice.