Mar 2019 – Jan 2022 · 2 yrs 10 mos · Bengaluru Area, India

• Bing Ads

Jun 2018 – Mar 2019 · 9 mos · Bengaluru Area, India

• Mainly worked on solving unstructured problems on text and Image data by leveraging deep learning and Machine learning techniques.

Jun 2015 – May 2018 · 2 yrs 11 mos · Gurgaon, India

• Worked as an individual contributor in Data Scientist role to build Deep learning algorithms to personalize offer content on Amex digital assets.
• Recurrent Neural Network - It has been used extensively in time sequence modelling. RNN models the offer interaction and context interaction in offer Recommender engine.
• Convolution Neural Network - the feature learned from convolution layer and max pooling layers will be fed into a fully connected layer and output layer to output the probability like prediction for each CM and each offer
• Word2Vec : In order to capture the Costumer offer interactions (clicks) as time series, used the event vectors with word2vec technique, such that offers which are clicked together are represented closer in vector space.
• FFM : This technique uses C++ based LibFFM package for field aware factorisation machine. It’s an effective method for classifying large sparse data including those from CTR prediction.
• XgBoost/LightGBM/CatBoost : Explored these gradient boosting techniques to generate features for RNN and FFM which were adding an additional value in the performance

May 2014 – Jul 2014 · 2 mos · Greater San Diego Area

• Objective: To infer differential signal in Genome data using statistical model
• Proposed a method based on the negative binomial distribution, with variance and mean linked by local regression
• Examined statistical testing like ANOVA test and Median polishing algorithm to decide difference in read counts
• Obtained an equation to compare different genome sequences allowing for more accurate comparisons
• Provided an effective algorithm for fitting the model to data, and show that it provides better fits

May 2013 – Jul 2013 · 2 mos · Delhi

• Market Price Cleaning application and Outlier Detection Algorithm
• Designed a MATLAB based application capable of pulling prices from different data sources, market feed, excel file, xml, MS SQL db
• Detect outliers in raw pricing data using regression analysis followed by employing methods Studentised algorithm and Cook's Algorithm and export clean data set in any desired format and display it graphically.