Main page
Hello there! I’m Alexander Wen, but just call me Alex. I am an alumni of the UBC MDS program, and you can expect this website about my work and tangibly related competitive programming antics to be updated on a roughly monthly basis.
Featured Projects
cdc_diabetes_prediction - Basic data analysis on predicting diabetes based on health and lifestyle features following sound data scientific workflows as part of the main project for DSCI 522 (Data Science Workflows), a course in the Master of Data Science program at the University of British Columbia.
imitation_game - imitation-game is a Python utility package for secure message processing. It provides a high-level interface for both Symmetric (shared secret) and Asymmetric (Public/Private key) encryption.
Package link - https://test.pypi.org/project/imitation-game/
Salescope - Salescope is an interactive dashboard that helps e-commerce businesses understand their customers better and reduce churn.
Link to the dashboard - https://019c8d13-5610-dd58-6134-331453179c0e.share.connect.posit.cloud
Amazon Patio, Lawn & Garden Search - LLM powered product recommender utilizing the 2023 Amazon Review Dataset for patio, lawn, and garden items. There are three retriever methods implemented in this project: a BM25 keyword retriever using DuckDB FTS index, a semantic retriever using FAISS and sentence-transformers, and a hybrid retriever combining both BM25 and semantic. These retrievers are supported by the usage of the qwen3-32b LLM model accesible through Groq for a RAG querying process.
Main branch deployment - https://amazon-recommender-alx649-joelyp.streamlit.app
Demo Video - https://www.youtube.com/watch?v=hO5vYL_v1gM&feature=youtu.be
Image Forensics & Fraud Detection AI Prototype
- Masters of Data Science Capstone Collaboration with GOcxm
- Constructed a fully reproducible multimodal machine learning pipeline for flagging of suspicious retail display images and metadata with Amanpreet Binepal, Oswin Gan, and Norton Yu
- Researched and implemented algorithmic improvements for similarity detection tools and module aggregation to retain detection of most fraud cases while reducing false positive rate by 40%
- Documented findings through automated experiments and reproducible Quarto reports
Competitive Programming Profiles
Here is a series of links to my main competitive programming profiles, ordered roughly in descending order of activity.
HonestCode: A Blog About Competitive Programming
A weekly blog for me to occasionally talk about competitive programming topics as a meaningful entirely human effort in this world where AI is becoming increasingly integrated into computer science as a whole for both good and bad reasons. Some of the general topics include:
- Actual algorithm and data structure explanations
- Contest recaps
- Practice methods
- A divulge of my thoughts on AI in the context of competitive programming that is only slightly less Um_nik-like
- Interesting problems