No More “It Works on My Machine”: Packaging & Deploying Production ML

​This session is part of the GirlsWhoML × Mentor Me Collective Workshop Wednesdays series.

​A machine learning model working in a notebook is only the beginning. In this session, Ipshita Chatterjee will show how ML models move from experiments to production — covering packaging, deployment workflows, CI/CD protection gates, safe rollout strategies, and real-world reliability.

​We’ll explore what happens after a model works locally, and what teams need to consider before shipping ML systems into production environments.

​Ipshita Chatterjee is a Senior AI/ML Research Engineer at Amazon and Head of Learning at GirlsWhoML, with experience across production ML systems, LLM workflows, and research-to-production engineering.

​This session is especially useful for students, early-career ML engineers, data scientists, AI builders, and anyone who wants to understand what production ML looks like beyond notebooks and demos.

​You’ll learn:

• ​how to package ML models for production

• ​why local environments often break in real systems

• ​how CI/CD gates help protect production ML workflows

• ​how safe rollout strategies like canary deployments work

• ​what early-career ML practitioners should know about MLOps