Skip to content

AI development · Archived resource

Build your first AI application

Explore a developer-focused introduction to Microsoft AI tools, infrastructure, services, and the ONNX ecosystem for intelligent applications.

2 min read Updated 25 Aug 2026
Cover of A Developer's Guide to Building an AI Application

This page records an ebook that introduced developers to Microsoft’s AI platform. It also covered supporting cloud services and tools for adding AI capabilities to an application.

The original form was imported without a working download asset. We have removed those inert fields rather than imply that a submission will deliver the ebook.

What the guide set out to explain

The ebook covered four practical questions:

  • how cloud infrastructure, data, and AI services combine in an application
  • which Microsoft tools supported model development and deployment at the time
  • how developers could add language, vision, or bot capabilities
  • where the Open Neural Network Exchange, or ONNX, could support model portability

Those topics remain relevant. The product names, interfaces, supported models, and deployment guidance have changed since 2020, so current Microsoft documentation should take precedence.

The story behind the resource

The original promotion opened with the work of Melisha Ghimere, whose student team designed an early-warning system for livestock health. The prototype used observations such as temperature, movement, sleep, and stress indicators to identify animals that might need attention.

Microsoft’s The Future Computed: Artificial Intelligence and its role in society described the project and the team’s participation in the 2016 Imagine Cup. It was included here as an example of developers starting with a defined human problem rather than a model or platform.

We have removed the imported accuracy figure because this archive does not retain the study design or test evidence needed to assess it.

Use current guidance for a new build

A modern AI application plan should begin with the decision or task, the source data, and the people affected. Then define:

  1. the model or service boundary
  2. the data and identity controls
  3. the evaluation set and acceptance thresholds
  4. human review and escalation
  5. monitoring, cost, and change ownership

That sequence is more durable than any list of product names. If the original ebook becomes available again with a verified download and reuse rights, this archive can link to it directly.

Continue reading

Related perspectives

Turn the article into a practical next step

Bring the use case, constraints, and current platform. We can help you identify what to test or decide next.