Hi, I'm Keilor 👋

I love using data to solve hard problems. I'm a team player, and I get things done.

I'm currently looking for new opportunities in Analytics & Data Science. If you're interested in working together, shoot me an email – I'd love to chat!

From 2021-2023 I worked in Revenue Analytics at Affirm, managing a client-facing team of analysts that oversaw revenue growth for half the business. Prior to that, I worked at Visa in Analytics and as a Chief of Staff, doing everything from Data Science to Sales Ops to Organizational Strategy. Before Visa, I built the underwriting and pricing models for CommonBond's student loan products, helping thousands across the country afford higher education. In another life, I was an options trader at Susquehanna International Group, and I studied Math at Yale.

In my free time I love to hone my technical skills by working on passion projects in Python, SQL, and machine learning. You can see some of my projects below!

Projects

Python

I explored defaults in LendingClub’s historical loan data and developed a strategy for using various tree-based models to mitigate investment loss. View project

I visualized the effects of the coronavirus pandemic on several technology and retail stocks and calculated the risk of investment. View project

I developed an advanced model to identify credit card fraud by comparing the performance of different machine learning models, neural networks, and sampling techniques. View project

I explored historical housing data in Ames, Iowa, and developed ensemble regression models to predict final sale price. View project

SQL

I became obsessed with this website called SQLZoo and solved all 270 SQL challenges on the platform. You can see my solutions in this GitHub repository.

I worked my way through all of the SQL challenges on HackerRank. Here are my solutions by difficulty — Easy, Medium, Hard.

I tackled all questions in the Udemy course 101 SQL Challenge Questions, using the AdventureWorks sample database from Microsoft. You can find my solutions here.