Sai Prakash R

ML Software Engineer II

Verizon Communications•  July 2022 - August 2022

 Dockerized NLP Models for easy deployment in production.  Built a Weekly Email report of varied metrics in Progression and Regression defects classified by the Log Analyzer Model.  Analyzed Jira defects data to determine the cause and reduce turnaround times.  Implemented SQL solutions to provide model performance metrics for dashboard.

ML Software Engineer

Verizon Communications•  August 2020 - June 2022

 Built an Image processing API that detects text, extracts it from an image, performs a spell check, and highlights the contour in the image if there is a spelling mistake.  Optimized and deployed multiple NLP classification models. Performed hyper-parameter tuning using Grid-Search. Evaluated using k-fold cross validation and deployed a Jenkins job that sends a validation report weekly.  Built a UI using streamlit that extracts fields from the Jira dashboard, predicts the label using the ML model, and updates the label in Jira.

Student Intern

Verizon Communications•  February 2020 - May 2020

 Developed a Machine Learning model that analyzes defect summary and description and assigns a technician for a bug or issue automatically.  Implemented Natural Language Processing techniques to extract information from text fields and identify the issue.  Performed Data Cleaning and used Smote for under-sampling of imbalanced data.

Data Science Intern

Flipkart•  May 2019 - July 2019

 Designed an NLP Model on unstructured review of the customer which predicts the topic and severity of the issue.  Built a Machine Learning model which predicts reviews over 7 topics and outputs the top 2 topics. Obtained an accuracy of 92%.  Implemented Vader Sentiment Analyser severity used to identify interventions for very bad reviews and alert the 'CSAT' team and also designed the presentation of the project to stakeholders