Skip to main content
Training programmeCompleted · as of 25 July 2026

AI Project Building Course — Cohort 9

An in-person, applied AI course that takes participants from a problem to a working prototype using CRISP-DM.

Context

Cohort 9 of the AI Project Building Course was a three-day, in-person programme that NLAI delivered in a training partnership with EMC (Educational Mastar Central). International Al-Ghurmani Company covered 50% of each participant's fee. The course project was a solar-panel inspection prototype that runs in the browser.

The course is designed for professionals, entrepreneurs, students and teams who want to turn an AI idea into a practical project. No prior experience in AI, programming or data analysis is required.

The ninth cohort met in person in Oman over three days in July 2026. It was a paid course; EMC (Educational Mastar Central), a Dutch training institute, is NLAI's training partner for the programme.

Challenge

Most AI courses stop at theory or at tool demos. Professionals, entrepreneurs and students who want to use AI in their work need to experience the whole path: choosing a problem worth solving, understanding the data it needs, building something that works and judging honestly whether it is good enough. The course had to make that path achievable in three days, including for people who had never written code.

NLAI role and contribution

NLAI designed and delivered the course, and ran its operations:

  • Curriculum — a three-day path from problem definition to a presented prototype, built on CRISP-DM.
  • Teaching — hands-on sessions that draw on the trainer's experience of previous AI projects, including the mistakes that usually cost teams time.
  • Course project — a working solar-panel inspection prototype, used to show every stage in practice.
  • Operations — online enrolment with a clear fee and consent step, and short feedback forms after each day.

Approach

The course follows CRISP-DM, a widely used method for data and AI projects, through six stages:

  1. Understand the problem and the project objective.
  2. Understand the data the project needs.
  3. Collect and prepare the data.
  4. Build the model or intelligent solution.
  5. Test and evaluate the results.
  6. Present the project and prepare it for use and further development.

Concepts are explained simply and applied immediately, and AI tools are used to build the project, so participants without a programming background can complete each stage.

Architecture

Input

  • Photos of solar panels
  • Live camera stream
Frames

Detection model

  • Compact object-detection modelYOLO
  • Exported for the browserONNX
Detections

In-browser inference

  • Runs on the device's GPU, or on its CPU as a fallbackWebGPU · WebAssembly
  • Images are analysed on the device
Findings

Inspection view

  • Dust, bird droppings, snow and damage marked on the image
  • Counts of affected panels and damage points
Architecture of the course project. It is a teaching prototype, not a production inspection system, and its accuracy has not been validated for operational use.

Implementation

  1. 01

    Day 1 — From problem to project

    Choose the problem and the target user. Define value, scope and success measures.

  2. 02

    Day 2 — Prototype and validate

    Build the core AI workflow. Test assumptions and improve the experience.

  3. 03

    Day 3 — Launch roadmap

    Plan data, operations, safety and measurement. Present the project and define next actions.

Outcome

  • Delivered in person over three days in July 2026.
  • A working solar-panel inspection prototype built as the course project, running entirely in the browser.
  • Participant feedback collected after each day through NLAI's feedback forms.

Verified figures

Enrolment applications recorded
27
Source: the programme's enrolment records, verified 18 September 2026

The figure counts every enrolment application recorded for the cohort, whatever its status: reserved, paid, cancelled or on the waiting list.

Technologies and methods

Technologies

  • YOLO
  • ONNX
  • ONNX Runtime Web
  • WebGPU
  • WebAssembly

Methods

  • CRISP-DM project method
  • Object detection
  • In-browser inference
  • AI-assisted building for non-programmers

What's next

Future cohorts are announced on the NLAI training pages. Organisations can also commission the same applied format for their own teams, built around a problem from their own operations.

Partners and relationship

Each partner is listed with its relationship to this work. None of them is presented as an NLAI client.

  • Educational Mastar Central (EMC)Training partner
  • International Al-Ghurmani CompanyCo-funder

Where this connects

The practice and capabilities this work draws on, for readers who want the technical depth.

Related pages
All work

Capability building

Planning applied AI training for your team?

We can shape the same problem-to-prototype format around a challenge from your own operations. Start with a short consultation about goals, audience and timing.