Aug 2021 – Jul 2024 · 2 yrs 11 mos · Sunnyvale, California, United States · Remote

• I work in a team of Data Scientists and ML Engineers that are key contributors in creating the automation and decision support platform, developing data processing pipelines, building AI/ML powered recommendation services to connect senior citizens to a fulfilling healthcare.
• Developed a Text Summary Application that converts a 25-page clinical PDF to a 2-page summary using Regex and NLP.
• This app extracts important clinical information thus reducing clinical review time by 8-10 min/doc.
• Built an Automated Workforce Planning Tool that efficiently allocates clinicians to facilities and identify optimal locations to hire.
• Workforce Planning Tool was used by Market Clinical Operations team for 34 states in US.
• Developed a Question Answering Chatbot to answer Patient’s questions related to Health Insurance Management.
• Designed and Built the Chatbot Architecture in RASA by defining intents, rules, policies and custom action servers.
• Evaluated Averbis product to check its feasibility in tagging incoming health documents to the existing patients in the database.
• Reported an increase in document tagging accuracy from 73.66% (Production Regex) to 90.22% (Averbis).

Jun 2020 – Aug 2020 · 2 mos · Philadelphia, Pennsylvania, United States

• Implemented “Orbis HR” to predict the sentiment of the employees based on their feedback reviews.
• Used Glove Embeddings and TF-IDF for feature extraction of web-scraped Glassdoor reviews. Performed training and validation on these features using Deep Learning models: RNN, GRU, LSTM, Bi-LSTM and transformer models: Roberta, Bert and Albert.
• Documented and Deployed a text annotation tool (Doccano) on our client data to test the performance of our trained models.
• Reported an increase in test accuracy from 69.68% (baseline using AFINN library) to 81.8% (using LSTM with Attention Module).
• Deployed the Minimum Viable Product on R Shiny by sourcing our python programs using Reticulate library.
• Published an article on “Introduction to Word Embeddings and its Applications” detailing Word2Vec, Glove, Elmo and Fasttext.

Jan 2020 – Dec 2020 · 11 mos · Philadelphia, Pennsylvania, United States

• Managed course content, conducted office hours to review student’s ML concepts and assignments.
• Prepared questions for quizzes and exam on topics ranging from Supervised Learning, Ensemble Methods to NLP.

May 2018 – Jul 2018 · 2 mos · Bengaluru Area, India

• Implemented “Employee Attrition Analytics” to predict the likelihood of employee leaving a company. Employed concepts such as Logistic Regression, Linear Regression and Correlation.
• Developed “Credit Risk Analytics” to build a model that can select or reject candidates that have applied for credit using concepts such as Logistic Regression, Decision Tree, Random Forest and Information Value.
• Forecasted Sales and Demand for a global superstore using concepts such as Arima Modelling and Coefficient of Variation.
• Technology used: R language.