Job Title: Data Scientist
Work Location: Tampa, Florida
Duration: 8+ Months
Job Description:
• Performed Topic Modeling for identifying abstract Topics underlying in the collection of J&J's covid vaccine customer complaints data using various algorithms including LDA, NMF, KMeans, etc.
• Created an API by utilizing Flask and imparted the plan to application group by enablin them to characterize the models.
• Utilized AWS Machine Translator for translating chatbot responses of employees from 10 different languages to English.
• Performed lexical normalization for translating/transforming a non-standard corpus to a standard text form.
• Used text simplification techniques (stopword removal, punctuation removal, lemmatization) for making a text easier to read and understand, while preserving its main ideas and approximate meaning.
• Extracted top keywords for identifying the most relevant terms to describe subject of the obtained intents from model.
• Created Text Summarization pipeline for shortened version of several documents belonging to the identified intents that preserves most of their meaning using cosine similarity an Transformers T5 Base models.
• Tagged entities in text with their corresponding type, typically in BIO notation using Named Entity Recognition techniques including StandcoreNLP, countvectorizer packages.
• Assisted in Entity Linking for the content strategy team by recognizing and disambiguating named entities and designing phrases, intents, and solutions for the knowledge base of J&J chatbot system.
• Predicted Demand for supply shipments of covid vaccines across the globe using supervised learning algorithms with an accuracy of 86%
• Built Phrases Generation pipeline using n-grams, fuzzywuzzy, sentence similarity models for identifying the top common phrases belonging to their corresponding topics.
• Developed Intent Detection tool using traditional NLP methods for capturing the semantics of chatbot transcripts of employees in the payroll division of J&J and assigning them to the correct label.
Work Location: Tampa, Florida
Duration: 8+ Months
Job Description:
• Performed Topic Modeling for identifying abstract Topics underlying in the collection of J&J's covid vaccine customer complaints data using various algorithms including LDA, NMF, KMeans, etc.
• Created an API by utilizing Flask and imparted the plan to application group by enablin them to characterize the models.
• Utilized AWS Machine Translator for translating chatbot responses of employees from 10 different languages to English.
• Performed lexical normalization for translating/transforming a non-standard corpus to a standard text form.
• Used text simplification techniques (stopword removal, punctuation removal, lemmatization) for making a text easier to read and understand, while preserving its main ideas and approximate meaning.
• Extracted top keywords for identifying the most relevant terms to describe subject of the obtained intents from model.
• Created Text Summarization pipeline for shortened version of several documents belonging to the identified intents that preserves most of their meaning using cosine similarity an Transformers T5 Base models.
• Tagged entities in text with their corresponding type, typically in BIO notation using Named Entity Recognition techniques including StandcoreNLP, countvectorizer packages.
• Assisted in Entity Linking for the content strategy team by recognizing and disambiguating named entities and designing phrases, intents, and solutions for the knowledge base of J&J chatbot system.
• Predicted Demand for supply shipments of covid vaccines across the globe using supervised learning algorithms with an accuracy of 86%
• Built Phrases Generation pipeline using n-grams, fuzzywuzzy, sentence similarity models for identifying the top common phrases belonging to their corresponding topics.
• Developed Intent Detection tool using traditional NLP methods for capturing the semantics of chatbot transcripts of employees in the payroll division of J&J and assigning them to the correct label.