Bikram Pandit

I developed a surveillance application for Android. It can take videos, capture images, or record audio in disguise or immediately whenever needed. This app has over 200,000 downloads with over 500 daily active users. I actively maintain it, so the app has been republished.

A translator for adding different languages to the application. This project can be integrated into almost any application with a low-cost subscription. The advantage for a developer using this tool is rapid application development without relying on a third-party application that can be expensive, require its extension, and, therefore, challenging to integrate.

It is a web application enabling users to collaborate and write in the same place. ShareByte is an excellent tool for instantly sharing a document online as it does not require any login but only a five-digit permalink. It synchronizes immediately among any number of sessions around the world.

A progressive web app that can be opened in the browser and installed as an application on Windows, Mac, Linux, Android, and iOS devices. This app is used in a pathology laboratory to record tests and print them in the required format.

I developed a late fusion neural network that can classify the mood of a song based on audio and lyrics. It is useful for a recommendation system such as Spotify or YouTube, where songs are targeted to the audience according to their mood, thus increasing user engagement.

I built a self-balancing platform using reinforcement learning and the PPO algorithm. It is better than a PID or model-based balancing system due to the development's complexity and the learning-based approach's robustness. It can be used in robot applications for balancing on uneven terrain or transportation.