PGIMER Chandigarh

Design, develop, and optimize deep neural network (DNN) models for automatic dysarthria detection, subtype classification, and severity grading using multi-modal speech data (acoustic, physiological, aerodynamic). Implement feature extraction, data preprocessing, and augmentation pipelines (MFCCs, jitter/shimmer, EGG, nasalance, etc.). Perform model training, validation (10-fold CV), transfer learning (e.g., wav2vec 2.0), and interpretability analysis (SHAP). Integrate DNN outputs into a conversational agent (CA) for personalized speech therapy exercises. Collaborate with speech-language pathologists for data collection, clinical validation, and benchmarking against FDA ratings. Document findings, prepare reports, and contribute to publications/manuscripts. Job Types: Full-time, Contractual / Temporary Contract length: 34 months Pay: From ₹80,000.00 per month Education: Bachelor's (Preferred) Work Location: In person