About

Swapnil Soni is a Senior Data Scientist/Machine Learning Engineer at PwC with over 12 years of relevant experience. He holds a Master's degree in Computer Science from Wright State University, a Bachelor's degree in Information Technology from Jabalpur Engineering College, and a Diploma in Information Technology from Kalaniketan Polytechnic College. Swapnil is also a Nanodegree holder in Deep Learning from Udacity. Swapnil is a data scientist with a software engineering background, and he has experience in working on various aspects of machine learning, including Natural Language Processing, Classification, and Clustering. He is proficient in handling both real-time and batch data using state-of-the-art technologies such as Hadoop and Spark. Swapnil is passionate about new technologies and always delivers projects from end to end. In his current role at PwC, Swapnil serves as a Senior Data Scientist, where he applies his skills and knowledge to develop cutting-edge machine learning models. Prior to joining PwC, Swapnil worked as a Senior Data Scientist/Machine Learning Engineer at Quicken and Sovrn Holdings, Inc. At Quicken, Swapnil developed a deep learning model and deployed it in AWS Sagemaker to categorize transactions. He generated embeddings from merchant names and costs (character n-gram) and provided them as input to the model, which achieved over 90% accuracy. He also worked on proof-of-concepts for creating various segments such as life events changes and location-based, applying various algorithms like SSA (singular-spectrum analysis) and Seq-to-Seq (CRF). At Sovrn Holdings, Inc., Swapnil worked on product similarity, where he used Keras, Spark, Airflow, and Docker to find the same product from a different merchant. He also used Bi-LSTM and LSH (Locality Sensitive Hashing) to find a similar product, which helped to replace the affiliated link with a different merchant in case of low product prices or high EPC (earning per click). Swapnil also helped the team move the creation of publishers' reports from queries to Spark based on EMR cluster (adhoc) and orchestrated the whole pipeline in Apache Airflow, which saved over 40% revenue in infrastructure. Swapnil's tech stack includes Data Scientist, Research Scientist, ML, Spark, NLP, AWS, Hadoop, Sagemaker, n-grams, Apache Airflow, Keras, Docker, and infra. With his diverse