2026-07-16 –, S4
Deploying machine learning models should be predictable, repeatable, and auditable—not a leap of faith. This talk breaks the deployment lifecycle into clear, actionable steps and demonstrates pragmatic choices for containerization, serverless and platform-based deployments, monitoring, and scaling. It also introduces the Model Context Protocol MCP as a practical pattern for giving deployed models secure, auditable access to external data and tools. Attendees will leave with concrete patterns, a checklist to avoid common pitfalls, and reproducible workflows they can apply the next day to move models from experimentation into reliable production services.
Ariane Djeupang is a Microsoft MVP, an ML Engineer and Project Manager, passionate about fostering community growth and innovation in the tech industry. While she is the current appointed Chair for PyCon Africa, she is dedicated to promoting diversity and inclusion, actively participating in the Python | Django | PyLadies communities and other initiatives that empower underrepresented groups in technology.