Nlp Scientist
CurrentResponsible for enhancing the end-user experience by understanding the user’s intent and supporting meaningful scenarios on Microsoft’s smart assistant and search capabilities.• Led feature development on Cortana LU scenarios and schema support in collaboration with the partner skills team.• Accomplished model architecture enhancement (SVM – CRF – BERT n-layer model) by leveraging contextual features thereby improving the query Precision/Recall metrics.• Opportunity analysis and exploratory research in prompt engineering for LLMs.• Research on analyzing improvement opportunities of Office search metrics with several query alteration techniques such as stemming, lemmatization on Compliant Workspace. • Successfully developed and deployed Entity Tagger models for search ranker results relevance improvement based on personalized lexicons with evaluation on real search style queries based on characteristics of eyes-off unlabeled data using unsupervised evaluation techniques.