I'm a Data Science student at UCLA interested in using data to understand problems, find patterns, and build tools around what I learn.

I originally got into programming through game development. Since then, I've worked with Python, SQL, R, PostgreSQL, machine learning tools, APIs, and data visualization across projects involving fraud, semantic search, e-commerce behavior, banking analytics, and computer vision.

I tend to learn by building. Most of my projects start with a question I want answered or a problem I think could be handled better, then I figure out what I need along the way.

I'm currently looking for data science, analytics, machine learning, and data-focused software engineering opportunities.

Python · FastAPI · PostgreSQL · pgvector · Sentence Transformers

Built a semantic search tool for finding related GitHub bug reports instead of relying only on exact keyword matches. Indexed 11,336 GitHub issues and combined embedding-based similarity search with live GitHub issue search.

What I worked on:

• Built the PostgreSQL database and vector-search pipeline
• Generated sentence embeddings for issue text
• Used pgvector for similarity search
• Built the API with FastAPI
• Connected the search backend to a usable interface

PostgreSQL · SQL · Data Analysis

Analyzed 1,000,000 bank account applications to find patterns associated with fraudulent accounts.

Key result:

• Identified a high-risk rule combining credit risk score and payment type that captured 2,084 fraud cases, or about 18.9% of all fraud in the dataset

Analysis included:

• Fraud rates across payment types
• Credit-risk segmentation
• Multi-factor risk combinations
• High-risk rule evaluation

Python · NumPy · Computer Vision · Model Evaluation

Working on methods for tracking biological cells across microscopy frames. This project is where I'm pushing further into computer vision, segmentation, tracking, and evaluation workflows.

Current focus:

• Building reproducible evaluation pipelines
• Comparing tracking approaches
• Debugging failure cases across image sequences
• Improving experimental structure and validation

R · Data Analysis · Visualization

Analyzed 500,000+ customer interactions to understand how users moved through an online shopping funnel and where the largest drop-offs occurred.

Focused on:

• Funnel conversion
• User drop-off behavior
• Stage-by-stage comparisons
• Data visualization in R

Excel · Power Query · PivotTables · VBA

Built an automated banking MIS workbook using customer, account, transaction, and card data.

The goal was to turn a large raw-data workbook into a reporting system that could refresh and update without manually rebuilding the analysis each time.

Included reporting for:

• Total transaction value
• Transaction volume
• Fraud rate
• Loans
• Support tickets

Languages: Python · SQL · R · Java · Lua · JavaScript
Data: Pandas · NumPy · scikit-learn · Excel · Power Query
Databases: PostgreSQL · pgvector · MySQL · SQLite
Backend / Tools: FastAPI · Git · GitHub · Docker · Linux · Bash · VS Code

2026 — Present

I work on technical benchmark tasks used to evaluate how AI systems handle software and data problems.

• Build Python and shell-based benchmark tasks
• Write automated checks to verify solutions
• Use Docker to make evaluation environments reproducible
• Debug evaluator failures and edge cases
• Work on database, SQL, ML infrastructure, and software engineering tasks
• Use Git and GitHub throughout the review and submission process

• Strengthening my statistics and machine learning fundamentals
• Building more end-to-end data science projects with model evaluation
• Improving my computer vision work through the cell-tracking competition
• Exploring quantitative research and data-focused engineering

I like building things that force me to learn something I didn't know before.