I am a Data Scientist at Dropbox with Product Safety Detection & Response Team, working on abuse prevention & cybersecurity.
I am interested in large scale data mining applied to Cyber-Security, Computer Vision & Healthcare. From 2012-17 I was a PhD candidate in Information Science (currently on leave, pending dissertation & defense) at Cornell Tech, Cornell University advised by Prof. Ramin Zabih. Instead of focusing on individual problems in particular domain, I am interested in building a new medium that has potential to enable new research paradigms & practical applications.
My approach to research is exemplified by Deep Video Analytics a platform for Data-Centric Computer Vision which I am building since 2017. Deep Video Analytics was conceived out of observation that in spite of deep learning driven advances in Computer Vision, there were no systems for processing visual data. DVA aims to formalize representation and processing of Visual Data in the same way search engines and relational databases simplified processing of text and structured data. You can follow the development activity on my Github profile.
Prior to DVA, during 2012-16 I built Computational Healthcare. Computational Healthcare attempted to speed up medical research by simplifying use of large de-identified patient datasets. It allowed researchers to perform rapid exploration and create new study designs. Computational Healthcare was used by researchers at Weill Cornell Medical College and lead to several papers/abstracts published in medical journals.
How can we develop a data-centric approach to Computer Vision?
How can we leverage large de-identified patient databases?
Copyright Akshay Bhat, 2018. All rights reserved.