From Laptop to Cloud: Deploying an AI Model Without Losing Your Mind
Building an AI model on a laptop is often the easy part. The real challenge begins when that model needs to run reliably as a cloud-native service. Many teams discover that a model which works perfectly in a local environment starts to fail once it is containerised, deployed, and operated in Kubernetes.
This session focuses on the practical journey of deploying an AI model from local development to the cloud. Rather than covering AI theory or model training, the talk explores real-world deployment challenges such as environment and dependency mismatches, containerisation pitfalls, configuration and secrets management, resource constraints, scaling behaviour, and unexpected failures in Kubernetes.
Using a simple AI inference service as a case study, the session demonstrates how cloud-native principles can help bridge the gap between "it works on my laptop" and a reliable production deployment. The talk includes a practical walkthrough of packaging an AI model into a container, deploying it to Kubernetes, and observing how it behaves under real operating conditions.
The session is aimed at developers, DevOps engineers, and cloud practitioners who want to deploy AI workloads in the cloud without unnecessary complexity. No deep AI background is required.
