Anurag Hambir Email & Phone Number

Las Vegas, Nevada, United States

Bloomington, Indiana, United States

• Employed advanced large language models (BERT and RoBERTa) in conjunction with PyTorch (Python) on a 300,000-text data set to create precise contextual word embeddings for generating personalized recommendations• Designing and leading the development of Graph Neural Network-based Recommendation system development (GraphSAGE, GATConv), enabling personalized foundation suggestions for grant recipients, and promoting reciprocal recommendations

United States

• Worked on the development of a customer retention model using Python for a Fortune 500 banking and insurance client, boosting the F-1 score by 70% through the XGBoost model and genetic algorithm, resulting in projected savings of $269,000 from retained customers• Utilized SQL and Snowflake to add new features to the Attrition model• Optimized and functionalized LSTM model for predicting the percentage of clients who churn

India

• Managed various data projects involving large unstructured data sets, while offering guidance and mentorship to a team of engineers.• Designed and implemented ETL pipelines for crawling data from different sources and stored them in MongoDB(NoSQL) database.• Built and deployed a recommendation engine for the OTTPlay application.

Pune, Maharashtra

• Implemented a data pipeline using PySpark to extract and analyze sentiments and entities from clients’ email data, employing Stanford’s CoreNLP library and SpaCy. • Processed and analyzed email data stored in Hadoop, through PySpark and stored the results in Hive tables.• Improved the performance of the data pipeline by integrating Kafka, reducing the total processing time by 50%.

Master'S Degree, Data Science, 3.9/4.0

Bachelor Of Engineering, Computer Engineering