I am working under the guidance of Prof. S. Umesh on multilingual and low-resource Text-to-Speech (TTS) and Voice Conversion systems. Responsibilities include developing and optimizing diffusion-based TTS models for Indian languages as part of the Speech Technologies in Indian Languages project, analyzing and modifying the F5-TTS architecture for low-resource multilingual adaptation, and focusing on improving duration modeling, alignment stability, and overall system performance.

Worked on robust audio projects at the Speech Information Processing Lab, including Autotagging and Automatic Audio Captioning (AAC), aiming to enhance system performance and accuracy under the mentorship of Prof. K Sri Rama Murty. Designed and implemented real-time inference pipelines, optimizing deployment on ROS using Jetson Orin to ensure efficient and low-latency performance in practical applications. Focused on end-to-end system deployment, including testing, monitoring, and optimization of audio models, ensuring scalability and robustness for real-world scenarios.

Collected and preprocessed data from diverse online sources through web scraping, ensuring high accuracy and reliability for academic research under the guidance of Prof. Salman. Applied advanced data analysis techniques to identify underlying patterns, and utilized regression analysis to generate actionable insights, contributing meaningfully to research outcomes and academic advancements.

Worked as a hybrid team member at Study Comrade, an edtech startup incubated at IIIT RKV, while pursuing undergraduate studies. Mentored students in data analytics and front-end technology projects, providing guidance, doubt-clearing sessions, and hands-on project support. Managed student batches, maintained the platform's website for smooth functionality, and leveraged analytics to enhance learning outcomes, gaining valuable startup and remote work experience.