Full Stack Data Scientist

Japkeerat Singh  • Peerlist

Kaggle competition in which one had to build a machine learning model with extremely less data that doesn't overfit on the testing data which was 100 times bigger than the training data. Final ROC score: 0.858 Final rank: 504(top 22%)

Don't Overfit

A GUI interface(JavaFX) application through which one can access SQL databases without any explicit knowledge of SQL and can control data using plain English. With powerful machine learning algorithms at the backend with carefully organized

Natural Language Interface for Database Processing

A django based web application that aids a college liaison department during placement season

Trabajo

With a combination of Image Processing, Deep Learning, Networking, Electronics, and Robotics, we designed and created a module of a robot that could be used to effectively segregate solid waste. The final module was able to correctly segreg

WasteBot

With an aim to improve the accuracy the first version of WasteBot was getting, we tried multiple approaches, from the ones tested in various research papers to some new ones.

At the time of writing this project here, we were getting an acc