* Design and develop the pipeline of a virtual assistant chat bot that acts as a bridge to mitigate the gap for knowledge transfer between Cypress and Infineon employees after the Cypress acquisition by Infineon Technologies. * Improved the accuracy from 65% to 95% for classifying large textual documents into multiple classes where hyperparameter was tuned through grid search.

Took the ownership to build a tool for monitoring, analyzing, and predicting the market trend in a given period thereby reducing man-hours for the SMA team. Implemented NLP features such as inverse-indexing for faster keyword search, siamese BERT for semantic searching, abstractive summarization, textual context clustering and subsequent heading generation for the cluster, Textual entailment, and sentiment analysis of individual texts. Therefore, enabling in-depth insight into the factors driving any market query that the user searched. Created the tool using HTML, Javascript, bootstrap, and flask. Also, visualized the insights using JS graphs e.g. D3.js, and deployed it on openshift(kubernetes).

* Classified(Inception V3) and clustered(Variational Auto Encoder ) various labeled and unlabelled defect silicon wafer images thereby saving 150K Euros year on year basis. * Wrote python automation scripts that can extract the news articles weekly and perform data wrangling/engineering before dumping them into MySQL table.

* Built a smart-recruitment system to help Infineon reduce their man-hours in recruiting candidates * Implemented identification of duplicate questions (siamese BERT and cosine similarity) and classification of questions into their respective categories (TF-IDF vectors). * Designed the UI and the back end for the tool and scraped the data set for the tool from Indiabix.com.