Skip to content
arula

Arula Academy / Learning area 05

Building AI agents.

Build agents that use tools, coordinate tasks, critique their work, and improve through deliberate evaluation.

Explore the courses

For developers building agent systems4 courses & practical tracks

Field notes / 05
A question to carry

An agent needs more than a good prompt.

Give it tools, boundaries, and a way to evaluate its work.

ARULA / LEARNING IN PRACTICE

From assistance to orchestration.

01 / What learning should change

The same promise.
Applied to your practice.

Build understanding you can explain, a language you can share, and habits you can use without a prescribed prompt.

01

Shared intuition.

Recognize where agent patterns help and where they fail.

02

A common language.

Describe routing, orchestration, reflection, and tool boundaries.

03

Everyday habits.

Build evaluation and stopping conditions into the system.

The goal is independence. Take these habits into unfamiliar work. The decisions remain yours.

02 / Courses & practical tracks

Choose the work
you want to get better at.

These modules involve writing and testing code. Agent-building modules offer Python and Node.js tracks; consult each module’s setup guide before starting.

05.01 / Module

AI-assisted development

Apply structured prompting, critique code, and validate improvements.

In this track

CGSE module 1 · Java and Angular

View course materials Course guide & exercises on GitHub
05.02 / Module

Core agentic patterns

Build chains, routers, parallel checks, and an orchestrator for incident response.

In this track

CGSE module 2 · Python and Node.js

View course materials Course guide & exercises on GitHub
05.03 / Module

Reflection and self-improvement

Build critique-revise loops, rubric-based evaluation, and reflection memory.

In this track

CGSE module 3 · Python and Node.js

View course materials Course guide & exercises on GitHub
05.04 / Module

Tool use and integration

Build a tested tool-using agent with error handling, performance, and security boundaries.

In this track

CGSE module 4 · Python and Node.js

View course materials Course guide & exercises on GitHub

03 / Make it your practice

Carry the learning
into real work.

Use each exercise to develop a way of thinking you can return to when the task, the tools, or the context changes.

  1. 01

    Choose an agent pattern

  2. 02

    Build the control loop

  3. 03

    Evaluate and improve

Keep exploring

AI-assisted software engineering.

Understand, test, improve, and secure software. Keep intent clear and engineering decisions grounded in evidence.

Explore this area View all learning areas