Group project using tensorflow, pandas, and multithreading in python
Knock Unlock is an authentication program that authenticates users by comparing a
real-time recorded knock pattern to any pre-set secret pattern.
- Utilized autoencoders and spectrograms for feature engineering
- Implemented neural networks with tensorflow
- Designed a multithreaded program to run the model in real time
Individual project using javascript (svelte) and python
Cricket Stats is a website built using Svelte that uses advanced data parsing and analysis to provide detailed statistics for cricket matches.
- Utilizes raw ball-by-ball data from the last 10 years to generate match statistics
- Offers head-to-head score comparisons to aid in creating fantasy cricket teams
- Built with Svelte and powered by a Python backend for number crunching
Individual project using pandas, scipy in python
- Designed an algorithm to decide portfolio allocations of a user-defined set of
securities based on past volatility
- Maximized the risk-to-return (Sharpe) ratio of the portfolio using periodic weekly
rebalancing
- Written in python and hosted on Quantconnect (a data provider and algorithm hosting
service)