Mondly VR — AI-Driven Language Learning in Virtual Reality

Mondly VR is an immersive language learning platform designed to help users practice real spoken conversations in virtual environments.

Instead of scripted exercises, learners interact with AI-powered virtual characters that respond dynamically to natural speech, creating realistic, engaging language practice scenarios.

The experience is built for standalone VR and focuses on immersion, responsiveness, and conversational realism.

My Role

Senior Unity Developer

  • Designed and implemented the core AI conversation pipeline inside Unity

  • Integrated Speech-to-Text → LLM → Text-to-Speech flows for real-time dialogue

  • Built modular, extensible systems to support evolving conversational behavior

  • Optimized performance and latency for standalone VR headsets

  • Worked closely with product, design, and engineering teams to align technical solutions with learning goals

  • Contributed production features across multiple release cycles

Impact
  • Enabled free-form spoken conversations with virtual characters in VR

  • Demonstrated early production use of large language models inside a real-time Unity application

  • Improved immersion and engagement compared to scripted dialogue systems

  • Delivered a stable, scalable architecture suitable for live product deployment

Technologies

Engine & Language
Unity, C#

AI & Audio
Speech-to-Text, Large Language Models (GPT APIs), Text-to-Speech

XR & Platforms
Meta Quest (Standalone VR)

Systems & Tools
Real-time networking with Photon Fusion, Google Text-To-Speech, performance profiling, Git

External Link