Viswanath Chepuri

MSc graduate Data Scientist with expertise in development using Python, SQL, Computer Vision, handling data, experienced in developing innovative machine learning solutions. Adaptable and eager to learn, aiming to contribute in data, intelligence roles to enhance innovative technology driven solutions

Controlled Bias Image Generation in Generative Models | PyTorch, Python,DL  Aug 2022 – Oct 2022 

• Developed an innovative approach for incorporating controlled bias into class representation by designing a custom dataloader for generating images from the widely-used MNIST handwritten digits dataset
• Conducted research on the generation of images, comparing the properties of synthesized images between GANsand VAEs by performing a high accuracy 98% CNN image classifier. Utilized PyTorch Deep learning framework for its flexibilty
• Observed a correlation between the proportion of class imbalance in the images and the resulting distribution of generated images. Utilized metrics to quantify the extent of this relationship and improving the distribution of generated images

Python Backend Application for Lyft Car Maintainence | Python, TDD, UML, Git     Feb 2022 – April 2022 

• Developed a robust backend architecture for notifying Lyft cars about required servicing using Python scripting, OOP, and software design patterns, improving fleet management and maintenance efficiency
• Utilized UML diagrams to effectively communicate system design and collaborated with the development team using Git to create a modular and scalable solution in Python
• Ensured high-quality software delivery through thorough unit testing and the adoption of Test-Driven Development (TDD) for the addition of new features and functionalities

Digital Financial Fraud:Analysis of Payment Data with Machine Learning | Python,ML,EDA   Aug 2022 – Oct 2022 

• Analyzed a dataset on digital financial fraud to develop a predictive model that can identify potential fraud cases
• Overcame the challenge of the dataset of 6 million rows being highly imbalanced by conducting a detailed exploration and cleaning of the data, including identifying and correcting possible discrepancies in the dataset’s description
• Implemented cutting-edge machine learning methods, including feature engineering and extreme gradient-boosted decision trees, to develop a highly accurate predictive model
• Achieved an enhanced predictive power of 0.997, as measured by the area under the precision-recall curve, indicating a high level of accuracy in identifying potential fraud cases

BlogIt:Social Blogging Website GCP deployement | GCP, REST, Javascript,GIT Feb 2022 – April 2022 

• Developed a web-based social blogging platform using JavaScript, with MongoDB as the database
• Deployed on GCP Kubernetes container and Heroku, managing application requests through Postman API
• Utilized RESTful API architecture to facilitate seamless communication between frontend and backend components, enabling efficient data retrieval and updates
• Integrated CRUD API functionalities within the application and added JWT tokens to secure user login credentials through encryption.

Ethereum Transactions Data Analysis | Python, MapReduce, Pyspark, Hadoop, SQL  Feb 2022 – April 2022 

• Performed deep analysis on the raw high-volume data of Ethereum transactions. The goal of this analysis was to support decision-making and gain a deeper understanding of the Ethereum transaction data.
• Implemented MapReduce and Pyspark to perform operations such as top K, aggregation, and moving window techniques to extract relevant data for analysis. Performing multiple RDD transformation operations like join etc.

British Airways Customer Experience and Purchasing Behavior Insights | Python,Pandas,ML Feb 2023–March 2023 

• Collected and analyzed 3100+ Skytrax customer reviews using web scraping, NLP techniques to gain insights into areas of improvement for British Airways services and enhance customer satisfaction
• Conducted a comprehensive study of customer buying behavior, developing a predictive model using the Xgboost classifier, VADER algorithm with an 85% test accuracy
• Identified key factors affecting British Airways’ service quality, such as food, inflight experience, and value for money, as well as the top features influencing customer purchasing decisions
• Provided data-driven recommendations to improve British Airways’ services and tailor marketing strategies, resulting in increased customer satisfaction, loyalty, and retention