Vaibhav Saini



          

Significant projects


1. HashPay (A Demo Payment system using Hashgraph) 

A demo payment system using python implementation of Hashgraph consensus algorithm. The application is still not stable for large number of users.
More information about it can be found here.

2. ERC20 wallet 

The wallet supports creation, storage and transfer of any ERC20 token. You can also deploy smart contracts using this wallet.

3. Implementation of Alpha Go for Chain Reaction 

Implementation of Alpha Go's algorithm (Monte Carlo tree search and neural networks) using Tensorflow on a self made desktop version game

4. Authox PLatform(Patent Pending) 

Developing a protocol for inter-blockchain identity verification and testing it on ethereum blockchain. Also(in progress) integrating the protocol to hardware wallet(made on arduino from scratch) for safely storing the keys.


Other projects


1. Stock Prediction using tensorflow 

Using Regression forecasting and predicting the stock prices for multiple stocks for different time periods.

2. Using Reinforcement Learning to train atari game BreakOut 

Using a different version of Q-Learning that is somewhat different from the original DQN implementation by Google DeepMind.

3. Using NLP and training different bots on Facebook & Slack 

4. Text detection in images using CNN(convolutional neural networks) 

Starting with MNIST dataset and testing it with different parameters
  • batch size
  • number of layers in neural network
  • different ratios of training cases:testing cases

Further using STN-OCR which is a DNN(deep Neural Network) to consist of a spatial transformer network — which is used to detected text regions in images, and a text recognition network — which recognizes the textual content of the identified text regions.

credits: https://medium.com/@Synced

5. Movie search engine 

Developed a movie search engine which just needs the name of the movie and searches the download link for the movie on web. No movie was hosted by me, instead all the content was searched over the internet and the content was served without any ADs.