I am the Founding ML Researcher at Verda, where I develop the machine learning behind content provenance. I built Verda's watermarking systems across image, video, and audio: imperceptible marks that survive the re-encoding of real-world distribution, so that a creator's work carries proof of its origin wherever it travels. As generated media becomes cheap and indistinguishable from the real thing, I posit that proving where content came from will become one of the internet's defining problems.
Alongside this, I collaborate with researchers affiliated with Stanford, Berkeley, and USC on making AI models better at understanding mathematics.

​

Earlier, I was a researcher at the Geophysical Inversion Group, Indian Institute of Science, where I developed deep learning methods for unsupervised and scarce-label settings, including a collaboration with ISRO to identify subtle elemental signatures in Chandrayaan-2 lunar X-ray spectroscopy. My work there led to my first journal paper, in IEEE JSTARS. I was advised by Dr. Pawan Bharadwaj, whose mentorship has been pivotal in shaping how I think about research.

​​​

​​I hold an integrated master's degree in applied geophysics from IIT Dhanbad. My time there exposed me to diverse fields, but it was a rekindled passion for research that ultimately defined my path. The thread running through all of it is representation learning: finding the structure underneath complex data, whether that is a watermark hidden in audio or a mineral signature hidden in X-ray spectra.

Beyond research, I am an avid quizzer and have represented my college at inter-IIT events. Music, particularly progressive rock, metal, and jazz fusion, is a personal passion, and I remain curious about the whys of life and what it means to be human.
Life may be too short to explore every idea, but connecting with fellow curious minds makes the journey richer.