Kevin Madura
Bio
Building real-world AI solutions for enterprise clients using DSPy, RLMs, and agent-native architectures. Technologist & expert witness @ AlixPartners.
Session (1)
Kevin Madura is a Director at AlixPartners who has built hands-on expertise applying DSPy and Recursive Language Models (RLMs) to real enterprise problems—structured extraction, codebase auditing, and agentic data analysis—and spoke at the AI Engineer Code Summit about why DSPy makes traditional prompt engineering obsolete. Attendees of his session get concrete, benchmarked techniques for making LLM-powered systems rigorous and production-ready.
Recent talks (1)
GitHub
@kmad- kmad/aie★ 92AI Engineer Workshop - DSPy is All You Need
- kmad/dabench-rlm-eval★ 22Benchmark harness for evaluating DSPy RLMs on data analysis tasks (InfiAgent-DABench)
- kmad/llm-structure★ 12A plugin for llm to support structured outputs
- kmad/rlm★ 6An implementation of the Recursive Language Models concept
- kmad/dspy-optimizer-experiment★ 3Using a dataset from Cleanlab.ai, evaluate improvements from using DSPy + GEPA optimization
Recent writing (5)
- A Data Scientist RLM That Lives in Your Program · blog · Mar 2026
- Auditing a Codebase for 87 cents in 50 lines of code using RLMs · blog · Jan 2026
- AI's Hedonic Treadmill vs. Task Horizon Exponentials · blog · Dec 2025
- Achieving 20 percentage-point improvement in structured extraction tasks using DSPy and GEPA · blog · Dec 2025
- Using DSPy to Detect Document Boundaries · blog · Aug 2025
Public activity researched automatically · as of Jun 2026