Generative AI and AI Agents Practice Exercises: How to Build Skills Between Lessons
Watching lessons feels like progress; doing exercises is progress. The space between lessons is where skills in generative ai & ai agents are actually built — and most learners under-use it.
Here is how to work the exercises in Generative AI & AI Agents so the hours you invest compound instead of evaporating.
Key points
- •Attempt first, then read the solution — never the reverse.
- •Labs produce portfolio evidence; keep clean, redacted records.
- •Space exercise sessions across the week; do not batch them.
- •Let quiz misses point you to the exercises that matter.
1. Do the exercise before the solution
Every Erudex exercise is written to be attempted first. Attempting badly, then reading the worked solution, produces far more learning than reading the solution and nodding — the struggle is what encodes the method.
If you are stuck for more than ten minutes, write down precisely where you are stuck. That sentence is usually the exact concept the next lesson teaches.
2. Use labs to convert knowledge into action
The labs in Generative AI & AI Agents are designed to be completed in real tools — a small working project with tests, design notes and a clear explanation of technical choices. A completed lab record with objective, decisions and result is a portfolio artefact you can show an employer.
Never publish credentials or confidential data in lab work. Redact everything you would not show a stranger.
3. Space the practice, don't batch it
Three 20-minute exercise sessions across a week beat one 60-minute session. Spacing forces retrieval from memory, which is the mechanism that makes skills durable.
End every exercise session by attempting one item you got wrong previously, without notes. When that succeeds twice in a row, retire the item.
4. Let the quizzes steer the exercises
Lesson quizzes at a 80% gate do two jobs: they check understanding and they point at which exercises matter. A wrong quiz answer maps directly to a lesson concept — go find its exercise rather than rereading the lesson.
Between modules, mix exercises from earlier modules back in. Interleaving is what makes the final exam's randomised paper feel familiar rather than cruel.
Frequently asked questions
- How many exercises should I do per lesson?
- All of them — they are written to be short. The discipline that matters is attempting first and reviewing afterwards, not the count.
- What if the exercises feel too hard?
- Step back one lesson or one module. Difficulty that feels absolute usually means a missing foundation from earlier, which the exercises are designed to surface.
- Do exercises count toward the certificate?
- The certificate comes from the final exam, but exercises are how you get there. The quiz gate and exam both reward exactly the skills the exercises build.
- Should I redo exercises I already solved?
- Only the ones you previously got wrong — and only until you solve them cold twice. Redoing solved work inflates hours without adding learning.
Study it properly: Generative AI & AI Agents
Build production generative AI systems and autonomous agents.
- Generative AI & AI Agents Study Guide: Skills, Practice and a Realistic Learning Plan
- Generative AI and AI Agents: A Practical Guide to Production Systems
- Generative AI Course Guide: Careers, AI Agents and AWS Exam Preparation
- Free Generative AI and AI Agents Practice Tests: How to Use Your Results to Plan Learning