← All work
Prototype

Ru2En Interpreter

An iOS Russian–English voice interpreter for conversations that need to move naturally in both directions.

iOSSpeech recognitionText-to-speechAirPods audio routing

Problem

Real-time bilingual conversations are difficult when people must repeatedly choose a language, manage turns, and switch between listening and speaking.

Solution

A voice interpreter that supports Russian–English translation in both directions while detecting language and conversational turns automatically.

Architecture

The iOS prototype combines speech recognition, automatic language and turn detection, translation, text-to-speech, conversation history, and AirPods audio routing.

Technical decisions

  • Treat each conversation as bidirectional rather than a single-direction translation task.
  • Use automatic language and turn detection to reduce interaction overhead.
  • Support AirPods routing for more natural hands-free use.
  • Keep conversation history available for context and review.

Challenges

  • Balancing automatic behavior with the ambiguity of live conversation.
  • Keeping the audio experience coherent across recognition, translation, and speech output.

Results

Public performance metrics are not available for this project.

Current status

Prototype.