AI Specialization Program

Recommended Assessment Structure

Instead of exams, we will use a 4-layer evaluation system.

1️⃣ Weekly Engineering Tasks (Light Assessment)

Every week includes:

Practical coding tasks

Short design assignments

Small implementation exercises

Graded on:

Correctness

Clarity

Structure

Documentation

Purpose: Continuous accountability.

2️⃣ Mid-Program Technical Review (After Week 4)

Assessment Type: Live technical checkpoint.

Students must:

Explain ML pipeline

Interpret evaluation metrics

Defend a modeling choice

Fix a small bug or improve code

This ensures: They understand fundamentals before touching GenAI systems.

3️⃣ Capstone Evaluation (Primary Assessment)

Weeks 7–8 are the real test.

Graded across 6 dimensions:

Category

Weight

Problem Definition

10%

System Architecture

20%

Model Performance

20%

GenAI / RAG Integration

15%

Ethics & Bias Handling

15%

Documentation & Presentation

20%

Total = 100%

Minimum passing: 70%

Distinction: 85%+

This makes certification meaningful.

4️⃣ Final Technical Defense (Oral Review)

Students present their system and answer:

Why this model?

Why this metric?

What are limitations?

How would you scale it?

Where can it fail?

This separates:

Copy-paste builders

from AI Actual engineers.

🏅 Certification Structure Recommendation

None generic certificate:

🎖 Certified AI Systems Engineer (Level I)

Awarded if:

All weekly work completed

Capstone completed

Passed final defense

🥇 Certified AI Systems Engineer — Distinction

Awarded if:

85%+ capstone score

Strong technical defense

Clean architecture

Responsible AI practices

This makes the title powerful.

 included assessment.

But of

engineering-based.  project-driven. t defense-based.

If they survive it?

Then salute them 🫡

And the certificate

Powered by Soft AI Africa | Training the Next Generation of AI Leaders in Africa.