AgenticRAGFrameworks

Building and Evaluating Data Agents

by DeepLearning.AI (with Snowflake)

IntermediateCourseFree~2 hours, self-paced

Build a multi-agent data agent, then trace and evaluate it with LLM-as-a-judge.

Start LearningReviewed July 8, 2026

Overview

Building and Evaluating Data Agents is a ~2-hour (1h59m) DeepLearning.AI short course created in collaboration with Snowflake and taught by Anupam Datta (AI Research Lead) and Josh Reini (Developer Advocate). Across 8 video lessons, 5 code examples, and 1 graded assignment, you design a multi-agent workflow with planners, executors, and specialized sub-agents that connect to data sources, then add tracing and evaluation to make it reliable. The course teaches LLM-as-a-judge evaluation, goal-plan-action (GPA) alignment metrics such as plan quality and execution efficiency, and runtime evaluations used to optimize the agent. It uses LangGraph for orchestration and integrates web search and retrieval from structured and unstructured sources, and is free during the DeepLearning.AI learning-platform beta.

At a Glance

Topic
Agentic
Level
Intermediate
Format
Course
Cost
Free
Duration
~2 hours, self-paced
Provider
DeepLearning.AI (with Snowflake)
Hands-on
Yes — code/exercises
Certificate
None

What You’ll Learn

  • Design multi-agent data workflows with planners, executors, and specialized sub-agents
  • Add tracing and observability to visualize and debug each agent step
  • Evaluate agents with LLM-as-a-judge and goal-plan-action (GPA) alignment metrics
  • Measure plan quality and execution efficiency, then use evals to optimize the agent
  • Connect agents to structured and unstructured data plus web search with LangGraph

Highlights

  • Built with Snowflake and taught by its AI research and developer-relations leads
  • Practical, eval-first: 8 lessons, 5 runnable code examples, and a graded assignment
  • Free during the DeepLearning.AI learning-platform beta

Who It’s For

Best For

  • AI engineers building agents over enterprise or analytical data
  • Developers who want a repeatable way to evaluate agent quality
  • Teams adopting LangGraph for multi-agent workflows

Prerequisites

  • Intermediate Python
  • Basic familiarity with LLMs and agent concepts

FAQ

What is Building and Evaluating Data Agents?

A hands-on DeepLearning.AI short course, built with Snowflake, on constructing and rigorously evaluating agents that answer questions over structured and unstructured data. It is for AI engineers who need their data agents to be measurably correct, not just demo-ready.

Is Building and Evaluating Data Agents free?

Building and Evaluating Data Agents is free to access.

What level is Building and Evaluating Data Agents for?

Building and Evaluating Data Agents is aimed at a intermediate audience. Recommended background: Intermediate Python, Basic familiarity with LLMs and agent concepts.

How long does Building and Evaluating Data Agents take?

Expect roughly ~2 hours, self-paced. Most learners work through it at their own pace.

What will I learn from Building and Evaluating Data Agents?

You'll learn: Design multi-agent data workflows with planners, executors, and specialized sub-agents; Add tracing and observability to visualize and debug each agent step; Evaluate agents with LLM-as-a-judge and goal-plan-action (GPA) alignment metrics; Measure plan quality and execution efficiency, then use evals to optimize the agent; Connect agents to structured and unstructured data plus web search with LangGraph.

Topics

data-agentsagent-evalsllm-as-a-judgelanggraphmulti-agent