About Me

Hello! Thanks for visiting my personal website.

I currently lead a team in the Applied Research group at Gracenote, working on problems in the Computer Vision and Natural Language Processing space. I also worked on Music Information Retrieval and recommendation problems for a few years, following my Master's in Music Technology at Georgia Tech. At Gracenote, my team is responsible for the full life cycle of an ML project. We help identify problems suitable for ML solutions, work with editorial experts to create a dataset, build, evaluate and deploy ML solutions, while communicating results and findings to stakeholders around the company.

I’ve explored and contributed to many areas of the business, starting with metadata and image acquisition, image enhancement, automatic image cropping, video processing, podcasts discovery and recommendation, and automatic text generation. My work has led to 5 patents (more yet to be public), one publication (more in the works), and some new music genres added to the Gracenote taxonomy. I’ve developed expertise in tools to rapidly prototype ideas and deploy ML solutions across multiple domains (video, text, and audio). Over the years, I’ve also cultivated an interest in entrepreneurship, inspired by the success of ideas that I’ve spearheaded and pitched.

Of late, I've been passionate about Reproducibility in Machine Learning, helping to raise awareness and build a foundation for reproducible ML at Gracenote. I frequently review papers for many conferences in the Computer Vision and Audio Signal Processing space like CVPR, ICASSP, ACM-MM, ISMIR and WACV.