← back
DSPy: The End of Prompt Engineering - Kevin Madura, AlixPartners
Takeaway
DSPy lets you write LLM software as typed, modular Python programs that survive model swaps and can be auto-optimized rather than hand-prompted.
Summary
- Kevin Madura (technical consultant at AlixPartners) frames DSPy as a declarative framework for building modular software where LLMs are first-class citizens with typed inputs/outputs.
- Walkthrough use-cases include sentiment classification, multimodal PDF processing, web research agents, PDF boundary detection, and arbitrary-length summarization, plus optimization with GEPO.
- Key value: control flow and intent are model-portable, so the program survives model churn; convenience like JSON parsing comes free.
- Contrasts DSPy with LangChain (less in-your-way) and Pydantic-AI; pitch is replacing prompt-string tweaking with proper Python programs that are then optimizable.
dspyprompt-optimizationframeworks
Original description
Applications developed for the enterprise need to be rigorous, testable, and robust. The same is true for applications that use AI, but LLMs can make this challenging. In other words, you need to be able to program with LLMs, not just tweak prompts. In this talk we'll cover why DSPy really is all you need in building applications with LLMs. We'll dive into real-world examples where we have successfully automated manual work using an opinionated DSPy-first approach to structuring applications, covering everything from simple modules to using SoTA optimizers to measurably improve performance. https://x.com/kmad/ **Summary** Kevin Madura, a consultant at AlixPartners, argues that building robust enterprise AI applications requires shifting from brittle "prompt engineering" to "programming with LLMs" using **DSPy**. He contends that prompts should be treated as implementation details optimized by the system, while developers focus on defining typed interfaces (Signatures) and modular logic (Modules). The session moves from a conceptual overview of DSPy's primitives—Signatures, Modules, Adapters, and Optimizers—to a live code walkthrough. Madura demonstrates real-world use cases, including a complex pipeline that routes files by type (SEC filings vs. contracts) and a "boundary detector" that uses visual layout to segment legal documents. The talk concludes with a demonstration of how Optimizers (like MIPRO) can automatically tune these programs to outperform manual baselines, followed by a Q&A on production costs and feedback loops. **Timestamps** 00:00 Introduction & The Enterprise AI Challenge 07:12 The 6 Core Concepts of DSPy (Signatures, Modules, Adapters) 13:23 Deep Dive: Class-based vs. Shorthand Signatures 19:57 Adapters: Controlling the Prompt Format (JSON vs. BAML) 24:17 Optimizers: The "Killer Feature" for Transferability 31:08 Code Walkthrough: Setup & Model Mixing 36:24 Handling Documents: "Poor Man's RAG" with Attachments 42:10 Adapter Comparison: Improving Token Efficiency with BAML 47:20 Optimizers in Practice: Creating Datasets & Metrics 51:13 Complex Pipeline: Routing & Classifying Arbitrary Files 56:00 Advanced Use Case: PDF Boundary Detection via Visuals 01:01:22 Analyzing Optimization Results & The "DSPy Hub" Concept 01:09:02 Q&A: Handling Delayed Feedback & Online Learning 01:13:00 Conclusion