-
Book Overview & Buying
-
Table Of Contents
DeepSeek in Practice
By :
In this chapter, we took you from an initial idea to a fully working service that replaces boring smartwatch summaries with dynamic insights powered by DeepSeek. We began by showing how to quickly build a simple prototype using the DeepSeek API to meet our initial requirements. We then explored running models locally with distilled versions and built a lightweight CPU-based service that runs without GPUs.
Next, we focused on flexibility, demonstrating how libraries such as LiteLLM let you switch backend providers seamlessly without touching your application code. Finally, we scaled up by deploying LLMs on AWS using LMI containers and vLLM, so we can deploy larger models while staying in control.
By now, you have the toolkit to take any DeepSeek-powered application to production, whether you want to keep things small and efficient, run via an API, or deploy larger models in the cloud.
But many successful applications today are powered not by a single LLM, but by...