Zain's Portfolio

5:16 PM

Thu Mar 13

Click or Press a key to unlock

linkedin|github

Who is Zain?

my name is Zain Moustafa,

I'm a Software Engineer, Data Scientist, and Machine Learning Engineer!

I'm a third-year Computer Science student at the University of Toronto, specializing in Software Engineering and Machine Learning, with a minor in Math and Statistics. I'm currently interning at Nokia and learning a lot about databases, dashboards, REST APIs, and JS!
With over 10 years in game development, I've helped create top 50 Roblox games that attract 1.5 million daily players. My work has been featured in 72 YouTube videos, racking up an impressive 902 million views.
Along the way, I've partnered with brands like Maybelline, Warner Music Group, and the American Heart Association, sharpening my skills in Git, Lua, SCRUM, SQL, JavaScript, Rust, C#, and Python.
Looking ahead, I'm passionate about diving deeper into research in Machine Learning, the financial market, and backend systems. My ultimate goal? To publish impactful research and/or create software that truly makes a difference.
When I'm not working on backend systems, web dev, Nokia projects, or dabbling in ML, you'll probably find me coding up new game mechanics. Want to see what some might look like? Check out some of my projects HERE

Education

University of Toronto
Graduating 2026
Computer Science with a minor in Mathematics and Statistics
GPA 4.0/4.0
2x Dean's List, 2x UofT Scholar, Scholar Award ($100k CAD), Dr. James A. & Connie P. Dickson Scholarship.
Teaching Assistant: CSC236

Experience

Independent Contractor, Roblox
Software Engineer, Game Developer
June 2015 - Present
- Led teams to create games with Lua, Rojo, Knit, and built systems with React, Node.js, and PostgreSQL.
- Developed data pipelines, analytics dashboards, and A/B testing using AWS S3 to store and process data.
Nokia
Software Development Intern
April 2024 - Present
- Developed a Python/C++ REST API script on a Linux environment and reduced runtime by 95% through an optimized stack data structure, automating the deployment of input data to Splunk databases.
- Collaborated with stakeholders to build 6 low-latency dynamic webpages using SPL, HTML, and JavaScript, improving accessibility and readability of billions of data for 3+ teams.
- Trained a Machine Learning transformer-based large language model based on the T5 family using Python to perform time-series analysis and predict future network failures, achieving 80% accuracy.
- Built a Jenkins ETL data pipeline using Splunk and Jira REST APIs to parse and push ticket data to Splunk.
- Utilized JIRA, GitLab, and Mercurial in an Agile environment for project tracking and code reviews.
Pardee Lab, University of Toronto
Machine Learning Engineer
September 2023 - June 2024
- Predicted GFP bioluminescence using Python's Pandas, PyTorch Variational Autoencoders, Gaussian Naive Bayes, Gradient Boosting Classifiers, XGBoost, KNN, and Logistic Regression models.
- Pre-trained, fine-tuned, and QLora'd LLMs (ProtT5) to achieve accurate protein sequence predictions.

Technical Skills

Languages & Web

Tools & Platforms

Libraries

Projects

kubernetes client

Improved K8s Java API by adding flags for efficient resource deletion, reducing exception handling overhead.
Introduced unit tests with WireMock and JUnit to ensure reliable resource deletion, boosting API reliability.

JavaWireMockKubernetesJUnitDocker

portfolio

Engineered the React frontend, built Express APIs with MVC to get data from a NoSQL MongoDB cluster.
Deployed website on an AWS EC2 instance, automated building and deploying using GitHub actions, and automated API tests with Postman, reducing time taken to add portfolio entries by 80%.

JavaScriptReactExpress.jsNode.jsMongoDBAWS

utmist website (student organization)

Led the infrastructure team in overhauling the website by managing issues, pull requests, and weekly meetings.
Collaborated with graphic designers to translate design mockups into visual elements using TypeScript, Tailwind, CI/CD pipelines, Vercel hosting, and Next.js.

TypeScriptNext.jsVercelGitHTML

stock trading app

Created a live trading Java app for real-time stock transactions using APIs, with JUnit for 90% test coverage.

JavaUMLJUnit

open street map path finder

Created a memory-safe C app to find the shortest path between 2 streets using Dijkstra's algorithm on a graph.

CValgrindGCCGDB

custom unix shell

Coded a Unix-like shell in C that implements string parsing, variable assignment, piping, and redirection by forking processes and managing input/output stream pipes.

CBashValgrindGDBGCC

captcha solver

Solved captchas using Selenium and trained AlexNet recognition model. Achieved accurate image recognition.

PythonTensorFlowSelenium

stock price prediction using covid-19 data

Used Python to automate cleaning and preprocessing of around 10,000 datapoints.
Performed time-series analysis for forecasting pandemic effects on market trends using lag plots, autocorrelation, and linear interpolation.

PythonRPandasMatplotlibNumpy

studeasy | hackthemist hackathon 2nd place winner

Developed a student-oriented web interface using Python, Flask, HTML, and CSS to generate notes, create test questions, and summarize syllabuses/books, decreasing time spent manually searching by 80%.

PythonHTMLFlaskscikit-learnCSS

gesture arithmetic

Enabled gesture arithmetic with Google MediaPipe for real-time landmark extraction, achieving 99% accuracy.

PythonPyTorchsklearnOpenCVMatplotlibTensorBoardGoogle MediaPipe

Show Applications

Google Chrome

Who is Zain?

Trash

Contact Me