What you will learn
  • Choose a demonstrable first goal.
  • Sequence concepts and projects.
  • Assess readiness through checkpoints.
  • Create reproducible portfolio evidence.
  • Separate reading time from practice.

Before you begin

Know the main AI terms and why outputs need verification. Revisit other Fundamentals lessons when a checkpoint exposes a gap.

Choose a result you can demonstrate

Learn AI is too broad to tell you what to do Tuesday afternoon. Explain a model, build a small predictor, or label familiar objects in a camera image are clearer goals. Each creates something inspectable.

For study habits, start with Using AI and AI for Learning. For software, prioritize Python. For robotics, programming and evaluation still help before you add motors and connections.

You need not master the field before useful work. You should distinguish a copied demonstration from understanding. Predict an outcome, run a check, and explain a failure at each stage.

Understand and evaluate first

After AI01–AI09, distinguish models from applications, training from inference, and generated claims from evidence. Draw a familiar system's information flow and describe a failure at each stage.

USE01–USE04 teach context, constraints, and examples. Add USE09 whenever facts matter. Produce a one-page explanation from supplied notes, preserve the source, and record your corrections.

Check readiness by identifying unsupported sentences even when they sound natural. Practice short source-bound tasks before open-ended research if necessary. Evaluation transfers across the whole curriculum.

Make code visible

PY01–PY08 cover variables, conditions, loops, functions, collections, files, errors, and environments. Write small programs and run them frequently. AI can explain a traceback, but predict the repaired behavior before accepting a change.

Build a program analyzing a few clearly labeled simulated measurements. Preserve units and handle invalid inputs. Running once is insufficient; explain what should happen with missing data too.

Use PY09 when arrays and tables become useful. Learn libraries to solve current problems rather than collecting names. The Python tutorial is a reference while our Python sequence offers smaller beginner steps.

Learn from data honestly

ML01–ML05 introduce features, labels, loss, fitting, and data splits. ML04 supplies an executable predictor. Keep a baseline and a test set that does not guide choices. Explain error in original problem units.

ML06 introduces networks; ML07 and ML08 introduce vision and language. You can learn from small examples or pretrained models without training a huge network. Inspect errors and whether inputs match intended conditions.

Record problem, data source, assumptions, baseline, method, evaluation, limitations, and one improvement. A modest reproducible result teaches more than an impressive unexplained number.

Connect software and hardware

Arduino introduces microcontroller inputs and outputs. Raspberry Pi introduces Linux, Python, networking, and cameras. Their roles complement each other in a robot.

Follow sense, process, decide, act, feedback, repeat. Test subsystems separately. Camera recognition does not automatically produce reliable motor stopping; control needs explicit limits and communication-loss behavior.

Begin with displayed readings or bounded light outputs. Progress to movement after motor, power, and communication lessons. The integration series then supports your original capstone.

Use repeatable practice sessions

Alternate learning one concept, reproducing an example, and modifying an assumption. These can happen on separate days. Reading estimates exclude installation, debugging, and experimentation.

Keep a log of question, prediction, observed outcome, and next test. Review recurring mistakes. If you cannot explain a copied line or metric, revisit its prerequisite.

Before an original project, specify the user, inputs, outputs, success criteria, and uncertain-case behavior. This makes the roadmap a design habit rather than a list of pages to finish.

Important terms

Checkpoint
An observable readiness demonstration.
Baseline
A straightforward comparison method.
Reproducibility
Repeating a procedure from recorded inputs and settings.
Subsystem
A component tested separately within a system.
Portfolio
Projects preserved with explanations and evidence.
Specification
Written intended behavior and success criteria.

Mini project: Plan four checkpoints

  1. Choose study assistance, software, data analysis, or robotics.
  2. Select four lessons and state a demonstration after each.
  3. Reserve separate reading and practice time.
  4. Choose a checked explanation, program, evaluation table, or diagram to preserve.
  5. Finish with success criteria beyond the number of pages read.

Common mistakes and debugging

  • Studying every tool before building: select a small project.
  • Skipping errors and environments: learn foundations that prevent confusion.
  • Measuring only reading progress: require tests and explanations.
  • Adding motors too soon: verify individual hardware subsystems first.

Independent challenge

Specify a model-plus-sensor project on one page, including uncertainty behavior and evaluation, even before you can implement it.

Check your understanding: 10 questions

  1. What makes a goal actionable?

  2. Why keep a baseline?

  3. Does an 11-minute article imply an 11-minute build?

  4. What belongs in a learning log?

  5. What precedes subsystem integration?

  6. In your own words, what does “Checkpoint” mean?

  7. In your own words, what does “Baseline” mean?

  8. In your own words, what does “Reproducibility” mean?

  9. In your own words, what does “Subsystem” mean?

  10. In your own words, what does “Portfolio” mean?

Quiz answers

Reveal all 10 answers after your attempt
  1. A result you can demonstrate and evaluate.
  2. It shows improvement over a simpler approach.
  3. No. Reading and practice estimates are separate.
  4. Question, prediction, observed result, and next test.
  5. Independent testing, including limits and failures.
  6. An observable readiness demonstration.
  7. A straightforward comparison method.
  8. Repeating a procedure from recorded inputs and settings.
  9. A component tested separately within a system.
  10. Projects preserved with explanations and evidence.

Summary

Progress through concepts, verification, Python, data, models, and hardware. Demonstrable checkpoints turn guided examples into independent design skills.

Continue learning

Start USE01 to practice giving a clear task, or PY01 if you want to write programs next.

Choose a connected learning path

Sources and further reading

Prepared 2026-09-18. Draft — curriculum links and progression reviewed