I am a tech fanatic, first and foremost. I have studied for my Bachelor’s degree in Computer Engineering at HBKU, and have taken a handful of Electrical Engineering courses at Texas A&M along with some Computer Science courses at Carnegie Mellon. I have worked with several tech companies before pursuing an MSc in Computing and IT Management at Cardiff University.
I produced code in Python, C, Java, MATLAB, Verilog, VHDL and 8051 Assembly.
Cryptocurrency Investment Adviser (CIA) Overview
This program is meant to be used as a helper for investors when making decisions related to investments into cryptocurrencies. This program relies heavily on the current crypto trends and gives advice on whether to invest or withdraw cryptocurrencies into fiat currency. The method used to measure whether the trend is going upwards or downwards is by instantaneous gradient as well as best line fit of the past 24 hours of data. I was going to add prediction based on ML on top of regression but since scikit-learn had a massive impact on user experience (loading times), along with the volatile nature of cryptocurrencies, the scope at which the graph and data used in this program has been narrowed down to 1 week. Analysis was done using a modified version of the Ordinary Least Square (OLS) regression method.
This decision was taken with the compromise of data prediction accuracy in mind. But in the shorter term of 1 week, such differences in accuracy between the two data analysis approaches are marginal compared to the cost of a painful user experience.
This was my first big Python project, did it because it was relevant to my work in the trading field at the time. Oh, and it coincided with the crypto bubble!
Zamanya: Segmentation, clustering and pattern analysis for sleep data.
Zamanya is an eHealth ML package that uses sleep data and analyzes datasets to identify good and bad sleep segments using a trained model.
I’ve spent my time during the summer going through an internship program with Qatar Computing Research Institute (QCRI). I worked mostly on the backend developing the preprocessing part of the package, including code that fetches raw sleep data via the Fitbit API and data mappers that output CSVs in the format usable by the package. This would help in further training the model using new datasets.
Because I’m bound by NDA, I am not allowed to share my work and it shall remain as proprietary information. References are available upon request.
Shopify Integration Gateway
A Flask-based RESTful API that allows Shopify merchants to directly interface with Snoonu's own Fleet Management System, FalconFlex. Allowing for Shopify order fulfillments to be directly sent to the FalconFlex system through HMAC-secured webhooks. Granular fulfillment progress updates are then sent back to the merchant through the gateway through another set of webhooks.
This was developed during my internship at Snoonu.
Source code bound by NDA.
Project HARPY
The project implements a monitoring system for the detection and classification of anomalous and rogue IoT devices within a local network using machine learning.
Developed data preprocessing functions including capture file processors and feature extraction algorithms for machine learning.
Computer Science Outreach
I have worked as a Development and Delivery Officer at Cardiff University’s School of Computer Science to develop and deliver various computer science materials and workshops to aspiring engineers of different ages and backgrounds.
Flask-based Dynamic Website
A Flask-based website that uses various web technologies including MySQL for data storage, OpenShift for deployment, Jinja for templating HTML pages, and Bootstrap 5 for styling.
Passport Visa Requirements Scraper
A Python program that uses MechanicalSoup (an extension of BeautifulSoup) to scrape passport visa requirements data from The Passport Index to output a CSV file containing each country's passport visa requirements for all the countries abroad.
This program was developed during my internship at sKora Tech.
Source code bound by NDA.
Wordle-solving Bot
A Python-based program that uses a dictionary of known words and the feedback of a word’s Wordle configuration (yellows/grays/greens) to give the best matching word for the current configuration.
Project PH03N1X: Prebuild (Timelapse Video)
Preparing a Threadripper-based system for a live building session. Got many more videos and documentation available upon request.
Proxy Server Setup (Presentation Video)
In this video, I explain the basic functions of a proxy server. FreeProxy was used to make a simple implementation for live demonstration.
