Yasin Patel

Computer Science @ Université de Montréal · Montréal, QC

I build full-stack web apps, ML systems, and developer tools.

01

Education

Université de Montréal

Bachelor of Arts and Science in Computer Science

Sep. 2024 — May 2027

Montréal, QC

Vanier College

DEC in Social Sciences & Math

Aug. 2021 — May 2024

Montréal, QC

02 · centerpiece

SkinFast

May 2026 · centerpiece

SkinFast is a real-time CS2 skin pricing and tradeup optimization tool. It scrapes the entire Steam Community Market, 21,000+ items, into a local SQLite cache, then runs an optimizer that finds the most cost-effective tradeup contract for any target skin. Under the hood it solves three problems: ingesting market data without getting rate-limited, modeling tradeup probability across multi-collection input pools, and ranking input combinations by cost-per-attempt and expected value.

Next.jsTypeScriptReactTailwind CSSNode.jsSQLiteREST APIs
SkinFast preview

01 — Web Scraper

Full Steam Market scraper with 21,000+ live items

A custom scraper crawls Steam's entire Community Market, 21,000+ items, with jittered delays, exponential backoff, and resumable offsets. A full pass completes in about 15 minutes without triggering rate limits.

02 — Real-Time Prices

Live pricing via stale-while-revalidate caching

Bulk scrapes every 4–6 hours hydrate a local SQLite cache, and on-demand refreshes keep high-traffic items current. Cached prices serve instantly while background jobs pull the latest data from Steam.

03 — Tradeup Algorithm

Probability modeling and cost-per-attempt ranking

For any target skin the optimizer enumerates input combinations across collections, models per-skin probability, and ranks candidates by cost-per-attempt and expected value, surfacing positive-EV contracts automatically.

04 — One of a Kind

The only site in the world with the most profitable tradeups

No other platform combines a full live-priced catalog with a probabilistic optimizer. SkinFast surfaces tradeup contracts that don't appear anywhere else, giving users an edge that simply isn't available on any competing tool.

03

Other projects

Rivals!

React Native · TypeScript · Supabase · Python · Riot Games API

Cross-platform mobile app for League of Legends stat comparison. BFS-based Python pipeline discovered 100K+ players across regions; custom impact-score ranking algorithm and database schema for auth, profiles, and match history.

Nov. 2025

ML Facial Expression Recognition

Python · TensorFlow · NumPy · Pandas

CNN-based facial expression recognition system trained on grayscale 48×48 images. Reached 80% accuracy through hyperparameter, batch size, and architecture tuning; automated inference and CSV export.

Oct. 2025

Roast-Me!

HTML · CSS · JavaScript · Python · OpenCV · Dlib

AI-powered web app delivering real-time personalized responses via computer vision and generative AI. OpenCV + Dlib facial-landmark pipeline feeds OpenAI prompts and ElevenLabs voice synthesis end-to-end in under 3 seconds.

Feb. 2025

04

Technical Skills

Languages
JavaScriptTypeScriptPythonJavaSQLHTML/CSS
Frameworks and Libraries
ReactNext.jsReact NativeNode.jsTailwind CSSTensorFlowNumPyPandasOpenCVDlib
Databases
PostgreSQLMySQLSQLiteSupabase
Developer Tools
GitGitHubDockerVS Codenpm
Concepts
REST APIsObject-Oriented ProgrammingData StructuresAlgorithmsWeb ScrapingAgile

05

Get in touch

Open to internships and full-time roles in software engineering. The fastest way to reach me is email.