Speaking-practice cards with pronunciation feedback
A language-learning app for listening to native audio, speaking along and reviewing pronunciation feedback. I built the app, API and web flow and corrected a case where answers were assessed against the wrong sentence.
Screens
A learning flow within each card
Cards combine words, sentences, examples and audio, linking listening and speaking to assessment results. Dictation, badges and saved cards are separate features, with progress tracked for each card.
Assessing the sentence actually recorded
A card contained both a main sentence and examples, so recording an example could be assessed against the main sentence. I passed the recording’s target sentence directly to assessment to align the input and evaluation reference.
From the app to service operation
I built the Flutter app, NestJS and TypeORM API, and website separately. My work also covered audio storage, OpenAI API integration, multilingual interfaces, server operation and acquisition and conversion analysis.
Architecture
