I am currently participating in a competition whose aim is to detect if a given text is generated by a LLM. My highest score is 0.940 AUC (current top 1 is at 0.964 AUC). I have made a demo application for my school teacher friend. The aim is to investigate if her students are using ChatGPT on their homework. Here is my running Github Repository for this competition. As the competition is currently ongoing, this is still work in progress.

I joined a competition that was hosted by the European Space Agency's Advanced Concepts Team where in I placed second in this leaderboard as of March 28, 2020. In this competition, we are given multiple low resolution images of each of 78 Earth locations and were tasked to develop an algorithm to fuse them together into a super-resolved image that is then checked against a high resolution image taken from the same satellite, PROBA-V. My solution can be found in this repository.

Implemented differential evolution algorithm to minimize a defined objective function. In this project, the algorithm was able to accurately model the frequency response of the system at 27, 30, and 34 Hz. Using linear regression did not accurately model the mentioned frequencies.

A collection of games wherein for every game an AI is trained and gets good with it! For training the AI, I used genetic algorithm and neural networks based on this paper. Also used reinforcement learning such as games like Pong. Github repository is found here.

A collection of algorithms visualized! I used d3.js and Python to visualize the algorithms. You can view the visualizations here.

Created a mini self-driving car powered by Raspberry Pi 4 and Coral Edge TPU. I also am planning to compare Coral and Neural Compute Stick 2. Github repository can be found here.