AISHWARYA SWETHA JONNALAGADDA

857-544-8161| [email protected] | https://www.linkedin.com/in/aishwaryaswetha/|

Boston, MA| https://github.com/AishwaryaSwetha
EDUCATION
Northeastern University, Boston, US Sep 2019-May 2021
Candidate for Master of Science (MS) in Data Analytics and Data Science GPA: 4
Mahatma Gandhi Institute of Technology, Hyderabad, India Aug 2014 -Mar 2018
Bachelor of Technology in Computer Science and Engineering
Received outstanding Academic Excellence award from JNTU-H
TECHNICAL SKILLS
 Programming languages: C, C++, Java, Python, R, MATLAB
 RDBMS/NOSQL: MySQL, PostgreSQL, Oracle dB, Neo4j, MongoDB, Bigtable, HBase, Dynamo DB
 Packages/Framework: Keras, TensorFlow, PyTorch, Scikit-learn, Spark, Django, Kubernetes, Docker, NumPy, Pandas
 Tools: Jupyter Notebook, R studio, Microsoft Office, GitHub, Google Colab, Tableau, Amazon Web Services, Google Cloud
Platform, Amazon EC2, Amazon Sage Maker, Flask
PROJECTS
Deep Learning using PyTorch Summer 2021
 Participated in a 6-Week course of Deep Learning with PyTorch, which includes coding, hackathon, Assignment each week.
 Implemented PyTorch Torch components, Melanoma Classification project using Resnet-50 with 54% accuracy.
 Protein Classification problem from Kaggle competition stood among top 15% of competitors.
Stock Price Prediction using Artificial Recurrent Neural Network -LSTM of Corporation Summer 2021
 Performed Time Series Analysis on Stock price attribute, namely ‘Close’ and Analyzed the trendline.
 Built LSTM Neural Network using Keras package and predicted stock price the following week or month accordingly.
 Model on the test data/new data has lower error rate of 5.31 (RMSE)Root Mean Square Error and hyperparameters like 80:20 split
of dataset,10-fold cross validations on the splits are used to evaluate and choose the best model performance on training.
Image Classification on CIFAR-10 using Tensor Flow Summer 2021
 Used CS-231n Stanford course instructions, built KNN, SoftMax classifier and SVM on CIFAR-10 dataset.
 Performed Logistic Regression, Random Forest, Convolutional Neural Network (CNN), with 4-layers CNN and 6-layer CNN.
 Accuracy of models have been increased from 38% (logistic regression) to 90% (CNN with n layer), CNN gives the best model
performance.
Big Mart Sales Prediction using Regression Models-Kaggle Spring 2021
 Performed Exploratory Data analysis, Data Preprocessing, Feature Engineering, Feature Transformation on Big mart Sales data
product wise and sales wise parameters reducing and manipulating 60% of overall inconsistent data.
 Applied Principal Component Analysis to find most contributors on Item_Outlet_Sales which was our Target Variable.
 Evaluated Accuracy using Lift Charts, Evaluation Metrics such as RMSE and compared each model fit, we improved our model
performance by 30% by using XGBoost Model which was suitable for our dataset.
Fake News Detection Using NLP Spring 2021
 Performed Text pre-processing techniques and NLP Features such as (TF-iDF, Bag of Words) is applied on data.
 Applied Logistic Regression on the data and required accuracy measure has been applied.
 Model is deployed using Flask components.
EXPERIENCE
Predictive Analyst -ANZ Virtual Internship May 2020-Jun 2020
 Performed Exploratory Data Analysis using Python on synthesized transaction dataset of three months transactions for 100
customers, Data Cleaning and Feature Engineering is performed with about 76% of improvement in data quality for model fitting.
 Insights have been made such as average transaction amount each month, spending’s of customers per week, Outlier Analysis
which distorts the analysis, visualizations have been made and documented.
 Implemented Predictive model suitable for the dataset, fit linear regression which gave poor results which 0.37% R-square
hyperparameters such as learning rate, number of features have been tuned with 0.45% R-square.
 Executed Decision Tree on this model for salary prediction which gave .78% R-square and correlation between variables have
been plotted.
Software Engineer -Automatic Data Processing (ADP), IND Mar 2018-Aug 2019
 Created self-service Analytical Dashboards which could help end user to investigate their payroll distributions
 Developed Python Scripts which used beautiful soup frameworks for evaluating web-scraping models
 Designed ad-hoc SQL queries for accessing and manipulating data on MySQL
 Performed Predictive Analytics on ADP workforce data predicting number of clients per business unit using Random Forest
 Analyzed Trends such as Overtime, Turnover, etc. on closely working with ADP Data Cloud Team
 Rewarded and Appreciated for being Best Innovative Employee among 20 chosen people under SBS-Business Unit