Engineer Assistant

Swetha Shiva Shankar Reddy

San Jose, CA ***** 469-***-**** ***************@*****.*** https://www.linkedin.com/in/swetha95/

SKILLS

Programming Languages: Python, SQL, C, C++, Matlab

Frameworks & Software: AWS Machine Learning, Azure Databricks, MapReduce, Postgres, Oracle, Hadoop, AWS Data Analytics, TensorFlow, TensorBoard, Keras, Pytorch, Detectron2, Microsoft Azure ML, AWS EC2

Machine Learning: NLP, Neural networks (CNN's & RNN's), Regression, Transformers, Random Forests Classifier, Clustering, Predictive Analytics, Pattern Recognition, LSH.

Libraries and Packages: NumPy, Pandas, Matplotlib, Scikit-learn, SciPy, Re, Imblearn, Keras

Visualization Tools: Matplotlib, Seaborn, Plotly, NetworkX, Pydot, Graphviz, Tableau, Power BI

WORK EXPERIENCE

Software Engineer, Capgemini, Bangalore, India July 2017 – Nov 2018

Vulnerability management using Tenable and Security Information and Events Management (SIEM) using QRadar and Splunk for the global education network Adtalem. Phishing email, Malware analysis and suspicious websites/URL/domain blocking on Cisco umbrella

Creation of dashboards and reports based on the customer requirements by collecting, analysing, authenticating and modelling Security metrics and Threat management data using the Business Intelligence (BI) tool Sisense. Security Operation Centre (SOC) monitoring and Incident response and management of SLA’s using the ITSM ticketing tool ServiceNow.

Instructional Student Assistant (ISA) January 2020 – December 2020

Worked as an Instructional Student Assistant (ISA) under Prof. Patricia Backer in the Technology department of San Jose State University to review Senior Design Projects and grade their work.

EDUCATION

MS in Computer Engineering January 2019 – December 2020

San Jose State University, San Jose, CA

Relevant Coursework: Data mining, Statistics for Data Analytics, Big Data, Large scale analytics, Machine learning, Deep learning, Cloud technologies, Data structures & Algorithms.

BE in Electronics and Communication August 2013 -June 2017

New Horizon College of Engineering, Bangalore, India

Relevant Coursework: Operating system, System software, Digital Image Processing

ACADEMIC PROJECTS

Text Summarization of Policy/Legal Documents Spring 2020 - Fall 2020

An NLP research project to implement an automatic text summarization using TensorFlow/Keras web plug-in by integrating Facebook’s User Preferences onto the Google Brain Research’s PEGASUS (Pre-training with Extracted Gap-sentences for Abstractive Summarization), a Transformer based model and evaluated the performance using the variants of ROUGE (Recall-Oriented Understudy for Gisting Evaluation) scores obtained for the generated summaries.

Performed fine-tuning on various parts of the network of PEGASUS and integrated user preferences such as length as an input to the summary to achieve summarization for “Terms of Use” Policies and based on the experiments, fine-tuning only the last layer of encoder and decoder yielded the best summaries with a ROUGE-1, ROUGE-2 scores of 0.6, 0.8 respectively.

Poker Game Classification Fall 2020

A Multivariate, Multiclass Classification problem in Python aimed to train a computer to classify the features extracted from a set of five cards into one of the 10 value categories, without the computer knowing the rules of the game.

Trained on a set of classifiers such as KNN, Decision Trees, SVM, Random forests, Extra trees classifier and performed voting to get the best F1-score to understand which model best fits the data and found that Random Forests performed the best with a Precision(weighted) of 0.59, Recall(weighted) of 0.62 giving a weighted F-1 score of 0.58.

Conference Paper Recommender system using Locality Sensitive Hashing (LSH) Spring 2020

Implemented a Content-Based recommender system in python for finding similar conference papers based off the title using Locality Sensitive Hashing (LSH) for nearest neighbors querying on a dataset of 1294 conference papers.

Performed Shingling, MinHashing and LSH to generate the signature matrix and calculated True and Estimated Jaccard Similarities to compare the results. Using LSH, achieved a speed-up of the recommendation process by 50 to 60%.

Classification, Detection and Assessment of Graffiti using Deep Neural Networks Fall 2019

Performed predictive analysis for the graffiti identification and detection by training object detection models using CNN algorithms such as R-CNN, Mask-CNN, SSD V2 and Detectron2 and achieved an overall accuracy, precision and recall of 0.7 to 0.8.

Created a dataset using a hashtag-based library Instaloader from which annotated xml files were generated by using LabelMe and LabelIMg too and a COCO dataset from json files for the Detectron2, a Facebook Pytorch Framework.

CERTIFICATIONS AND TRAINING

Azure Spark Databricks Essential Training, Tableau 10 Essential Training, Azure: Understanding the Big Picture, Python for Machine Learning and Data Science Bootcamp, Master SQL for Data Science, Learning MongoDB, Oracle Database 12c, AWS: Data Analytics, Neural network – Theory & Applications, Recommender Systems.

Contact this candidate