I’m currently a Computer Science Major at the University of Virginia, graduating in 2018. I have interned for IBM's Cloud department twice and Amazon's Recommendation Platform this past summer.
During the school year, I am heavily involved in Student Game Developers organization as both a Director and a Treasurer. For more information, check out my resume!
I am currently directing a team of 15 designers and programmers to make a multiplayer web-based game. Built with Unity and C#, Super Stellar is based off of io games such as Agar.io and Slither.io. Although the game is in its early stages right now, around December this game will be demoed to the rest of the Student Game Developers organization. So please comeback then to see the games progress!
DanCI
I worked with two other IBM interns to create a modified Travis CI prototype called DanCI. It contained all the functionality of Travis CI but was modularized so that 3rd party integrations can easily be added to the continuous test and deploy process. Using docker scripts and JavaScript, I designed and scripted the first modules that were integrated into DanCI.
Goal-ly
I worked with 3 other friends at Pearl Hacks to create a sleek and simple website for goal management. By filing out a simple Typeform in Goal-ly, the user can automatically view and update their goals as a blog. We used to AWS to deploy the website and setup mySQL databases and Zapier to automatically parse the data from Typeform.
Improving Recommendations in Transactional Emails
Transactional emails include the order confirmation and ship confirmation emails that you recieve when you purchase an item on Amazon. I worked on solving two problems in this space: improving coverage and filtering substitute items.
1.) Improving coverage: Recommendations are often based on the purchase history of a customer. However, for some customers, no recommendations are generated. This can occur because the customer is new and therefore does not have a purchase history. In addition, recommendations sent to customers via email must be appropriate. Recommendations for sensitive items (i.e. swimwear, alcohol, adult items)
2.) Filtering substitute items: One problem that customers face is receiving recommendations for items they just purchased. In the example above, I received recommendations for wallets when I had just purchased one. In order to solve this problem, I added a filter to emails that would rectify this problem.
About Me Without Me
I worked with 2 other friends to create About Me Without Me in a 24 hour hackathon. This application scrapes data from an individual's Facebook and Google data archives, using only unstructured data (search history, photos, messages, etc.) in order to create a structured profile that can be used for targeted marketing/advertising. Checkout the project here: https://github.com/dkaruturi/hackNC.
Ticket System
This is a tech support application that I created using Ajax, JQuery, DOM, PHP Mailer, and the Twilio Messaging API. It allows administrators to manage their tickets and customers to create new tickets. Click Here to Learn More: https://github.com/dkaruturi/TicketSystem
Rainy Days
Rainy Days is a 2D platform game about a child who has fallen into a world where it always rains. Armed with only an umbrella, the child must navigate the cloudy world, solve puzzles, and avoid nimboids to return home. I worked on level design and overall gameplay design. Check it out!
COOL Interpreter
A fully functional interpreter for C.O.O.L. (Classroom Object Oriented Language). Each of the four phases (Lexing, Parsing, Type-Checking, and Interpretation) is written in a different language. Click here for more details