Context Engineering with DSPy

Chapter 2. Introduction to DSPy

A note for Early Release readers

With Early Release ebooks, you get books in their earliest form—the author’s raw and unedited content as they write—so you can take advantage of these technologies long before the official release of these titles.

This will be the 2nd chapter of the final book. The GitHub repo will be made active later on.

If you’d like to be actively involved in reviewing and commenting on this draft, please reach out to the editor at [e-mail].

You now understand what context engineering is, and why it matters. You’ve seen the failure modes that plague long contexts, and techniques for fixing them. Now it’s time to explore DSPy’s core architecture, and how you can use it to solve context engineering problems. DSPy was designed from the very start to be portable, modular, and self-improving.

• Portable: programs can run across different models and providers.

• Modular: programs are composable, reusable, and easily modified.

• Self-improving: powerful optimizers that evaluate and improve the underlying prompts.

The key to DSPy is understanding the core primitives it offers for building AI programs. DSPy programs are built and optimized with five foundational concepts:

• Language Models: Model provider configuration and prompt formatting

• Signatures: Task definitions declaring inputs, outputs, and initial instructions

• Modules: Reusable components for building program pipelines

• Evaluation: Defining metrics and datasets for measuring ...