Info! Try clicking on the dots.
Menu

Mohammed Perves Student Engineer Strategist Visionary Avid   ML   Enthusiast Sports   ML   Enthusiast A   bit   of   an   athlete

Veritas | Lux | Scientia

Projects

×

InvestOpia

Python | Flask | MySQL | Yahoo Finance | scikit-learn | Tweepy | NLTK | News API

InvestOpia is an in-house data and analytics platform, created to serve the internal users at MPire Capital.

MPire Capital is a tiny quantitative hedge fund, that seeks to generate Alpha (excess risk adjusted - superior returns) by leveraging Analytics and Machine Learning (ML).

InvestOpia performs advanced analysis using the collected data such as sentiment and predictive analytics using:

Image Repo with content based search (object detection)

Python | Flask | MySQL | TensorFlow | OpenCV | cvlib

This is an image repo with the following features:

  • Content based search - you can search for images based on the objects or contents within the images
  • Metadata search - you can search for images based on the metadata (such as tags, description) associated with the images
  • Upload images - You can upload images to the repo
  • More features under development

The cvlib was used for object detection and all the data is stored on a MySQL DB.

Ryzer

Python | Flask | MySQL | AWS DynamoDB | AWS Lambda

Ryzer is a Web app that allows students to view and share resume templates, in hopes of bridging the network gap. Also, has a front-end for a resume parser API, to give users insights into how their resumes are parsed by an ATS.

Ryzer has served 1500+ unique students, from all over the world, extending from the University of British Columbia to Dalhousie University, all the way to UC Berkeley and MIT as of March 10, 2021

Azu: Quantitative Trading Strategies

Python | Jupyter Notebook | scikit-learn | Keras | NLTK |
Pairs Trading
  • Used a word to vector algorithm to represent stocks (a combination of features) as vectors
  • Used cointegration, cosine similarity and k-means clustering to find "pairs" and develop pairs trading strategy
Momentum Trading
  • Scraped social media sites like Reddit (Wallstreetbets) and Twitter to find optimal (hype) securities for momentum trading
  • Performed sentiment analysis on the scraped data and combined with technical indicators to perform momentum trading
Algho Trading
  • Currently building and training reinforcement learning trading agents (bots), based off research and alghorithms from OpenAI.

Stock Price Prediction

Python | scikit-learn | Tweepy | NLTK

Machine learning application created in Python, leveraging scikit-learn and yahoo-finance API, that uses support vector regression to predict stock prices. Also, includes twitter sentiment analysis for improved accuracy.

Source