ai/courses

Home / Providers / DeepLearning.AI

Evaluating and Improving Agentic Applications

The eval course most agent tutorials skip: traces, test sets, code evaluators and LLM judges, judge reliability, and turning production failures into regression tests. Take it once you have an agent worth measuring.

Go to the course on DeepLearning.AI
Provider
DeepLearning.AI (with Arize AI)
Price
Paid
Certificate
Paid certificateNeeds DeepLearning.AI Pro.
Level
Intermediate
Time
15h 7m
Format
Video course with code examples and graded assignments (5 modules)
Checked
2026-11-05

Graded assignments and certificate need DeepLearning.AI Pro; the page does not state a free audit option. Check current pricing.

01

What you will learn.

  • 01Tracing
  • 02Benchmarks and test sets
  • 03LLM judges
  • 04Tool-call evaluation
  • 05Regression tests
  • 06Phoenix

More inAI agents coursesEvals & testing courses

02

Before you enroll.

From the course page, checked 2026-11-05.

Who it is for
  • Anyone who builds AI agents or LLM apps and wants a reliable way to know if they work.
You need
  • Comfort reading and writing Python
  • Basic pandas
  • Prior experience calling an LLM API
What you build
  • An evaluation suite for a tool-using agent, built with Arize Phoenix
  • A reusable test set built from labeled agent failures