June 4, 2026 · 8 min read
5 Bad Reasons Teams Avoid DSPy
Developing LLM prompt-based tasks? DSPy can help you build systems that are easier to test, improve, and trust.
Applied AI Research Team Lead · Author · Educator
Design AI systems that improve from evidence.
I write and teach practical ways to evaluate, optimize, and ship reliable AI systems—so coding agents work inside systems we understand and control.
I lead an Applied AI Research team at ContentSquare, creating the culture and engineering environment teams need to win with data.
Featured course
Founding Student Early Access
Data-Driven AI Engineering
Build AI systems that improve from evidence.
A hands-on path for software engineers, data engineers, and data scientists who want practical experience building reliable AI systems.
- 01Build a working baseline
- 02Evaluate real behavior
- 03Improve from failure data
- 04Ship with confidence
See the course
The first two lessons are free.
Latest writing
Notes on building reliable AI systems
May 31, 2026 · 3 min read
What if your users aren’t “buyers” or “non-buyers”, but different kinds of shoppers moving at different speeds?
An ML question of the week based on unsupervised clickstream clustering for user behavior analysis.
May 21, 2026 · 4 min read
A new book is born: Building LLM Applications with DSPy
Why DSPy is becoming the PyTorch for reliable prompt-based tasks in the agentic AI era.
Published by Manning
Building LLM Applications with DSPy
Replacing manual prompts with systematic optimization
A practical guide to defining, evaluating, and optimizing reliable LLM applications—with an open repository of code and examples.
Explore the book