Portfolio
Speech & Language AI
Real-time pronunciation coaching across seven languages — built, deployed, and live right now, not just prototyped.
AI Pronunciation Coach
Whisper • Real-Time Audio • ML Scoring • MLflow
See It In Action
Score Trend Tracking
Every practice session tracked, so improvement is visible, not assumed.
Core Capabilities
Multilingual Practice
Practice sentence generation and coaching across German, French, Spanish, Portuguese, Russian, Japanese, and Chinese.
Specific, Actionable Coaching
Not a vague score — targeted notes like ‘keep vowel timing even’ or ‘focus on pitch consistency.
MLflow-Powered Evaluation
Scoring model performance tracked via MLflow, deployed through Docker on Render.
Technology Stack
AI Pronunciation Coach integrates real-time audio processing, speech AI, and experiment tracking into one system.
Frontend
- Streamlit
- Streamlit-mic-recorder
- Real-Time Audio Capture
Backend
- FastAPI
- Uvicorn
- Docker
AI / ML
- Whisper API
- scikit-learn
- NumPy
- SciPy
- pandas
MLOps
- MLflow
- Render Deployment
- Streamlit Cloud