Speech processing and representation learning

Absence of ground truth information makes elemental mapping of lunar surface very challenging, especially in the presence of varying solar flare intensity. To tackle this issue, I'm working on extending the application of SymAE in unsupervised scenarios on Chandrayaan-2 data. This collaboration with ISRO scientists is through IISc.

Member of Geophysical Inversion lab at the Centre for Earth Sciences. I research novel deep learning architectures and statistical learning methods with application in earth sciences.

Explored the application of symmetric autoencoders (SymAE) in hyperspectral imaging for remote sensing, resulting in a paper currently undergoing peer review. The preprint is accessible on arXiv at https://arxiv.org/abs/2309.14286

During the COVID-19 pandemic in 2021, I got inspired by startup culture from social media. With a friend, I started PPTHut, a service for designing PowerPoints and graphics. We aimed to help local teachers in southern Assam adapt to online teaching. My friend handled creativity, while I managed the website.Looking back, I realize I was quite naive during those times. This experience humbled me, teaching me that reality doesn't always align with our initial perceptions. It was a lesson in managing expectations and understanding that depth comes from diving headfirst into a subject, rather than observing from a distance. My view of the world shed its rosy tint, and I emerged wiser for it. In the process, I did learn a few things about design and its effects on human psychology, search engine optimization, and different marketing techniques.

Treasurer at EAGE student chapter of IIT(ISM) Dhanbad