Harsha Gowda

My name is Harsha, and I’m a graduate student in the Electrical and Computer Engineering department at the University of California, Davis. I build accessible technologies that empower human expression—focusing on language and speech modalities.

I develop fluid neuromuscular interfaces using EMG, MEG, and EEG signals, enabling seamless human-computer interaction through hand gestures, handwriting, and speech. My current work includes a wristband-based interface for decoding handwriting and gestures, and an orofacial EMG-based interface for reconstructing speech from silent articulations.

At UC Davis, I lead efforts to collect and open-source large-scale surface EMG datasets, involving hundreds of participants performing gestures, handwriting, and speech articulation. All our data and code are available to the research community to promote transparency, reproducibility, and collaborative progress. Learn more through my blog and publications.

My research explores the theoretical foundations of multivariate biosignals. I use geometric machine learning and graph neural networks to develop efficient abstractions of EMG, MEG, and EEG time series—models that can generalize across individuals and adapt with minimal data.

My broader vision is to leverage AI to augment human capability. I aim to create egocentric multimodal systems that enable humans to interact with technology as naturally and expressively as they would with another person.

Before UC Davis, I earned a Bachelor’s degree in Avionics from the Indian Institute of Space Science and Technology and worked at the Indian Space Research Organization , where I helped build and launch satellites .