Useful AI friends for developers – building a multilingual app

A word like 'Home' can break your app's translation. Discover how specialized AI tools solve this common internationalization pitfall.

#1about 2 minutes

Inspecting untrusted code with ChatGPT's code evaluator

Use ChatGPT's code evaluator to safely inspect, clean, and understand a zip file containing an unknown project before running it locally.

#2about 6 minutes

Refactoring a poor UI layout using Azure AI Studio

Provide a system prompt and detailed instructions to an AI model in Azure AI Studio to refactor and improve a poorly designed Windows Forms UI.

#3about 2 minutes

Renaming UI elements programmatically with GitHub Copilot

Use GitHub Copilot to refactor code by renaming generic UI elements like 'button1' to meaningful names derived from their hardcoded text labels.

#4about 3 minutes

Extracting hardcoded text for internationalization with AI

Prompt an AI model like ChatGPT to extract hardcoded language strings from source code and generate an i18n-compatible JSON resource file.

#5about 1 minute

Understanding the limits of general LLMs for translation

General-purpose LLMs can produce incorrect translations for ambiguous words like 'home' or 'resume' due to a lack of specific context.

#6about 5 minutes

Improving translation accuracy with context and glossaries

Use specialized translation tools like DeepL with context hints and custom glossaries to resolve ambiguity and enforce brand-specific terminology.

#7about 5 minutes

Automating the i18n process in a JavaScript app

Structure a front-end application using i18n keys and automate the translation of resource files with an API for a predictable and efficient workflow.

#8about 3 minutes

Q&A on AI code security and decoupling translations

The speakers address audience questions regarding the security of sharing code with AI services and best practices for decoupling language resources using i18n standards.