Shree Harsha Bokkahalli Satish

I'm interested in speech-to-speech conversational AI, signal processing and machine learning. For my PhD, I will be working on enchronic speech, which involves how quickly we respond in conversation, how we co-create meaning with our conversation partners, and other real-time interaction dynamics. Alongside this, I’m also researching bias detection and mitigation methods in conversational AI models.

Personal website: https://shreeharsha-bs.github.io/

I'm supervised by Éva Székely and Gustav Eje Henter. My research is supported by the WASP program.

Previously
I have a Master’s in Electrical Engineering (Signal Processing and Communication) from IIT Bombay, where I built automatic speech recognition (ASR) models for literacy assessments in India which is now in field use by an NGO. I later did an MSc in Data Science & AI to deepen my expertise in ML at the University of Liverpool.

I’ve also worked on LLM applications for the aerospace industry, where I built a RAG-based system that automated building new repair schemes as an Innovation Engineer.