Currently working with audio team - resposibilties include development of new features and performance improvements of the existing onesMulti-process for AISound app - -Reduced Proportional Set Size(PSS)+SWAP memory during boot time -Implemented intelligent mode features(Active Voice Amplifier and Adaptive Sound)-Refactored existing code base(C++) to use sub-process and used shared memory for inter-process communication-Reduced memory footprint of the application by almost 50 percent compared to previous version. Currently being used in samsung smart TVs worldwideSoundbar algorithms optimization-Used ARM NEON intrinsics(Single Instruction Multiple Data) for C code optimization of audio data processing. Reduced total MCPS by 48 percent-Working on Hifi4 DSP optimization(SIMD) for algorithms of soundbar using xtensa xplorer tool. -Task includes reducing MCPS count as a whole. Currently able to reduce the MCPS by 62 percent

Completed B.Tech from Indian Institute of Technology, Ropar

CUSTOMER CHURN PREDICTION - • Worked on identifying customers who are likely to drop off in the periodof next three months using historical data• Used state of the art Machine Learning(ML) algorithms such asAdaBoost, SVM with gaussian kernel for model building• Due to sparsity and data imbalance, built three different models fordifferent partitions of data. Combination of these returned highprecision and recall values on test dataSMART SEARCH TOOL FOR AUTOMATED COMMENT CATEGORIZATION -• Worked on a web-based application that was used to search phrases incustomer feedback database• Designed and implemented functionalities for phrase recommendationusing state of the art Natural Language Processing(NLP) techniqueswhich in turn reduced the manual efforts required for commentcategorization by more than 75%