Stack

  • Vue.js
  • Node.js
  • TypeScript
  • Jest

The Work

Apple’s chatbot teams had collected millions of conversations from 26 locales in 7 languages. I built tools that let them explore that data and classify utterances through both automation and human review.

One chat-history app was selected among the top 10 of 150 submissions at an internal data science conference. I also built a tool for reviewing model predictions and added Jest and ESLint to the development workflow.

Decision Give teams a way to inspect conversations and review both automated and human classifications.

Result Teams worldwide could review and classify conversations from 26 locales in 7 languages.

What I Did

  • Built a chat-history app selected in the top 10 of 150 internal submissions.
  • Helped teams explore conversations from 26 locales and 7 languages.
  • Built automated classification and human review workflows.
  • Built a separate tool for reviewing model predictions.
  • Added Jest and ESLint while following Apple’s design, security, and engineering standards.
  • Worked with business and data science teams from design through delivery.