We love to hack with new technologies and experiment on interesting idea.
An experiment to try out FFmpeg within the browser. It’s a ffmpeg.js powered purely frontend app to split any video into 15-second clips needed for Instagram stories. The app runs completely on the browser without any backend
A simple experiment to try out Rescript/Reasonml. Built for Freelancers and Software services company to calculate their (weekly/monthly/annual) earning based on their rates and availability in USD/INR
A decentralized application built on top of Ethereum using solidity to reduce email spam. The project allowed users to put a paywall in front of their inbox. The sender of an email can bid on several Ethereum tokens he is willing to give for the receiver to read his/her email. The money was held in a smart contract and only if the receiver reads the email, money would be transferred to his/her account otherwise after predetermined days, money would be sent back to the sender’s account.
This started as a project for one of the clients to scrape Linkedin and developed into a fun experiment where we are trying to build a CLI interface for Linkedin. We hate Linkedin UI/UX and wanted to get relevant information about potential customers/clients without navigating different tabs. The project also comprises a new simple data serialization framework.
An experiment to automatically extract slides from videos (tech talks, online lectures on youtube, etc). As a team, we consume a lot of tech talks and we often need to access a particular slide from some talk that we saw. To solve this problem, we are working on an ML system that can automatically identify slides within the video, extract those slides, deduplicate them and convert them into a PDF. Few research papers have talked about this problem, but there is no clear solution. We are experimenting with open cv and neural networks and heuristics right now.