With the given framework for the Neural Net, this project was about optimizing the convolution operations on the CUDA GPU using different parallel processing techniques.
As a tennis fan, I sought to understand the numbers behind one of the greatest tennis rivalries. I collected the data and cleaned it, and then used it to draw interesting charts and graphs, using Python and Matplotlib.
In this simulation, an autonomous car that is being controlled by a Deep Reinforcement Learning algorithm must weave its way through traffic at the greatest speed possible upto 80 mph. I optimized the hyperparameters to increase the speed of the vehicle from roughly 50 mph to 71 mph. This code is for the Deep Traffic competition hosted by MIT.
In my Computer Systems Engineering course, I built a simple kernel inspired by Linux on x86, in a team. We incorporated different features of the Linux kernel, such as the File System, RTC, Paging, and the Process-switching code. Additionally, we wrote the drivers for the keyboard and the file system.
Made in the summer before the POTUS elections of 2016, this project was about data visualization of the public sentiment of the political candidates. Using the Twitter Rest API, my team and I created interactive graphs by collecting sentiment values with the help of the HPE Haven OnDemand API.
This project won my team the Twitter prize at Angelhack Delhi, 2016.