Sonakshi Reddy Morramganti

Graduate Research Assistant

Mixil Lab, USC•  August 2019 - Present

• Research and implementation of denoising autoencoder for gap filling of soil moisture data that is used by NASA. • Upgraded database infrastructure to reduce redundancies and ensure scalability for new sensors with 50% faster insertion and querying time

Machine Learning Engineer I

Phenom People•  March 2017 - February 2019

• Enhanced user experience by 20% by improving personalized recommendations and search on different platforms. • Collaborated in end-to-end system design and architecture of ML features and coordinated with product design teams to ensure scalability and efficiency of machine learning infrastructure. • Optimized deployed ML models' performance by regular monitoring and evaluation of logs and data input. • Designed and implemented scalable and personalized recommender system by developing hybrid algorithms of topic modeling and clustering on millions of data points. • Developed Natural Language Understanding (NLU) search feature using NLP techniques (Entity recognition and extraction) and deep learning (Seq2Seq models) to understand user's intent and provide relevant search results. • Built knowledge graphs using NLP techniques and data processing technologies (Spark, Kafka, and Neo4j) to improve quality of recommendations and search results ranking. • Prepared clean and standardized data using NLP techniques and regular expressions. • Reduced DB retrieval time by optimizing queries and modifying elastic search schemas for various indices. • Mentored intern to build Named Entity Recognition (NER) model to improve feature engineering process. • Guided 4 member team to build universal skills representation model to use it in different downstream applications.

Data Science Intern

Phenom People•  December 2016 - March 2017

• Improved search results ranking by incorporating machine learning models (Semantic Search) and tuning search relevance score with better query augmentation which increased customer engagement by 10% • Maintained product deployments in Jenkins, elastic search clusters in production, and other search-related product requirements in production through CI/CD to ensure consistent and accurate results.

Data Science Intern

International school of Engineering•  July 2016 - December 2016

Improved inventory estimation of large automotive parts manufacturing company by implementing inventory forecasting model using ARIMA, Neural Networks, and Moving Averages.

University of Southern California (USC), Los Angeles

Computer Science, MS•  May 2019 - Present

GPA: 3.78 Courses: Natural Language Processing, Machine Learning, Artificial Intelligence, Analysis of Algorithms, Database Systems, Geospatial Information Management

JNTU, Hyderabad (Jawaharlal Nehru Technological University)

electronics and computers engineering, B.Tech•  August 2011 - May 2015

Graduated with 81.12% (First class with distinction) and was in the top 15% of the class Relevant Courses: Object-oriented programming in Java & C++, Data Structures and Algorithms, Databases