Tech-lead in Instagram Ads Delivery, Ranking, Quality and Supply Team. Leading a team of more than 10 engineers on various projects ranging from:1) Ads supply personalization: Personalize ads supply in instagram by balancing the tradeoff between revenue and user sentiment with the app. Figure out ways of ranking ads within organic content of the user to increase overall user ads satisfaction.2) Ads ranking for long term revenue gains: Most of the ads ranking is done for realizing short term ads conversions. We are working on modifying ads ranking value function to increase long term ads engagement of the user.3) Ads quality: Include various factors like ads negative feedback, survey based modeling, etc. to prevent bad ads experiences.Besides these, I have worked on a lot of ranking problems for instagram in the past ranging from ads ranking, shopping ranking in feed and stories, generating unconnected recommendable shoppable content for the user, etc.

Guide: Prof. Arvind Narayanan, Computer Science, Princeton University1) Extracted more than 400 semantic and syntactic features from data logged by OpenWPM, a tool whichcrawls websites in a humanly fashion, logging data like http requests and response headers, cookies set, etc.2) Used various machine learning models like LSTMs based deep recurrent neural networks, SVMs and random forests for classifying more than 10 million unique URLs as tracking or non-tracking.3) Achieved a cross validation accuracy of more than 99% with a false positive rate of less than 0.5%.

1) Worked in Ads Personalization team to reduce ad user language mismatch on Facebook.2) Trained models to learn based on ad conversion types, user country-language pairs and other features to get empirical 0.5% increment in revenue and 40% decrement in language mismatched ads.3) Acknowledged for my work by receiving a full-time job offer from Facebook.

1) Developed code to establish robust, low latency exchange connections for high frequency trading.2) Ensured quick processing of enormous live market data by efficient algorithms and compiler optimizations.

1) Developed a J2ME application for streaming audio data over HTTP on non-RTSP supporting feature phones.2) Work resulted in a Research Paper at ACM Conference and an award at mBillionth Awards South Asia 2013, given for innovations in mobile and telecom domains.