KRITHIKA SRINIVASAN Email : KRITHIKASRINIVASAN.0@GMAIL.COM GITHUB.COM/KRITHIKA-SRINIVASAN Linkedin : LINKEDIN.COM/IN/KRITHIKA-SRINIVASAN Education • State University of New York at Buffalo Buffalo, NY Buffalo, NY Master of Science in Computer Science (GPA –> 3.4/4.0) August 2018 – December 2019 • Mumbai University Mumbai, India Bachelor of Engineering in Information Technology August 2014 – July. 2018 Experience • Oneture Technologies Mumbai India Data Scientist April 2020 - Present ◦ Content Summarizer : Used pretrained BERT Huggingface encoders and PyTorch to implement an abstractive text summarization solution that generates two-line summaries of descriptions of 150-word texts with an accuracy of 92 percent ◦ Recommendation Engine : Developed a content recommendation engine using Python’s SpaCY library and fastText embeddings to dynamically and accurately display similar frameworks ◦ Automated Risk Assessment : Developed a risk assessment model framework in Python with Scikit-Learn for a client in the cash logistics sector which classi es employees based on the risk of cash shortages they might be responsible for. This solution has achieved an F1 score of 0.90 ◦ Mission Scorecard : Performed data transformations using Apache Spark in Azure Databricks to aggregate data across the Azure data platform for 36 monthly KPI metrics • PatternEx San Jose, CA Software Engineering Intern June 2019 - August 2019 ◦ Model Validation : Added a new validation component to the code base to evaluate 80 network threat classi cation models against a golden dataset ◦ Unit Testing Framework Development : Created a unit testing compoment with more than 200 unit tests using Python’s PyTest framework and Apache Spark to test the data preprocessing ow in detail • Atos Mumbai, India Data Science Intern June 2017 - August 2017 ◦ Customer Feedback Analytics Tool : Created an interactive tool in R Shiny that perform text analytics on client feedback extracted from SQL Databases using R’s tidyR libraries for wrangling text data to generate real-time results using R’s ggplot2 package ◦ Employee Turnover Predictor : Created a model that uses ARIMA to generate a model from time-series data to predict employee turnover on a monthly basis. The model performed with an overall accuracy of 88 percent Projects • Master’s Research Project - Uncovering Bias in Twitter bios (R, Pyspark) : 600,000 Twitter bios along with the users’ gender, age and political a liation were imported using Pyspark. By performing two kinds of short text topic modelling, Scholar and Biterm Text Modelling (BTM) in R, 15 distinct groups and words strongly associated with these groups were identi ed The output was a list of words with their classi cation based on their associations along gender and political lines • Explainable AI Versus Neural Networks (Python, Keras) : Developed a supervised learning model using Random Forests and a deep learning model using Keras to determine if two images of handwriting samples have the same author. Obtained 78 percent accuracy with the Random Forests model, and a 95 percent accuracy with the Deep Learning model • Bachelor’s Research Project - Real Time Twitter Analytics (R, R Shiny) : Developed an application that captures tweets about a user-speci ed topic in a time-frame set by the user and then performs analytics on them in real-time and presents the results interactively. Analytics include statistics about the captured tweets, three kinds of sentiment analysis and a network of interrelated words Programming Skills • Languages : Python, C++, SQL, R, Scala Technologies : Azure, AWS, Tensor ow, Keras, NLTK, Spark, Docker