Research Scientist at AI4Bharat, IIT Madras, specializing in Natural Language Processing and Text-to-Speech systems. Published researcher at top-tier venues (NeurIPS, Interspeech) with expertise in building large-scale multilingual AI systems for Indian languages. BS in Data Science (IIT Madras, 9.52 CGPA) with a unique interdisciplinary background combining Philosophy and Computer Science.
Research Highlights
RASMALAI
Interspeech 2025
Introduces RASMALAI, a large multilingual dataset for controllable text-to-speech, and develops IndicParlerTTS, a controllable multilingual text-to-speech model for 24 languages.
The State Of TTS: A Case Study with Human Fooling Rates
Interspeech 2025
Introduces Human Fooling Rate (HFR) to directly measure speech indistinguishability, revealing that CMOS-based parity claims are misleading via HFR evaluation of TTS systems.
IndicVoices-R: Unlocking a Massive Multilingual Multi-speaker Speech Corpus for Scaling Indian TTS
NeurIPS 2024
Created large IndicVoices-R dataset for multilingual Indian text-to-speech. Benchmark evaluates zero-shot Indian TTS, model improves generalization across diverse speakers and languages.
Featured Projects
Smriti - Cultural Heritage Data Platform Full-stack platform for collecting India’s largest multimodal cultural dataset covering all 7,000 taluks. Built Flutter mobile app, Flask backend, and Vue.js viewer for offline-first data collection.
Indic Parler-TTS Multilingual text-to-speech model for 24 languages with fine-grained speech control. Curated 8,385-hour dataset and contributed to training setup and research publication.
Work Experience
AI4Bharat, IIT Madras
Research Scientist | January 2025 - Present
Data Science Intern | February 2024 - December 2024
- Contributing to the TTS Lab by cleaning, preparing data, and training models
- Building open-source text-to-speech models for the 22 official languages of India
- Co-authored multiple research papers published at NeurIPS and Interspeech
IIT Madras
ML Research Intern | September 2023 - January 2024
- Developed content bridging theory and practice in classical machine learning
- Created high-quality ML notebooks exploring model internals and guiding effective EDA
Teaching Assistant - Machine Learning Techniques (CS2007) | May 2023 - August 2023
- Prepared comprehensive course notes adopted as standard course material
- Mentored students and conducted sessions on classical ML topics
Education
| Degree | Institution | CGPA | Year |
|---|---|---|---|
| BS in Data Science and Applications | IIT Madras | 9.52/10 | 2025 |
| BA in Philosophy | Christ University, Bengaluru | 9.57/10 | 2021 |
Technical Skills
Research Areas: Text-to-Speech, Natural Language Processing, Large Language Models, Multilingual AI, Speech Synthesis
Languages: Python, JavaScript, SQL, Java
ML/DL Frameworks: PyTorch, TensorFlow, HuggingFace Transformers, Scikit-learn, Pandas, OpenCV
Web & Backend: Flask, Django, Vue.js, SQLAlchemy, Celery, Redis
Mobile Development: Flutter, Dart
Tools & Infrastructure: PostgreSQL, MinIO, Docker, Git
NLP/LLM Tools: llama.cpp, Langchain, Llama, Mistral, Bloom, Flan-T5
Certifications
Deep Learning Specialization - Coursera Mastered CNNs, RNNs, LSTMs, Transformers, and optimization techniques
Generative AI with Large Language Models - Coursera LLM lifecycle, training, fine-tuning, and deployment strategies
