Chanyang (Judah) Ryoo

Often CCAR and CECL purpose models are grouped together, and they have to do with models designed to satisfy regulatory and accounting standards required for banks. Here I briefly describe their differences.

CCAR is a US regulatory standard that requires banks to conduct stress tests under the scenarios that the Fed provide. These scenarios have 9 quarters of duration. The stress tests show whether the bank has sufficient capital reserve under the scenarios.

CECL is an international accounting standard that requires banks to calculate expected lifetime loss of loans and set aside appropriate reserves.

For both exercises, credit risk models and economic scenario forecasting models are used, as these work together to provide estimates of losses.

If you are an applied mathematics PhD and looking for jobs, I would recommend looking into model risk roles at a bank. These jobs are also called model validator roles. I would say they are the “academic” jobs in a corporate world.

The main role for these jobs is to look at the conceptual soundness of models used by banks and validate them, challenge them, stretch them, etc. I found this was kind of similar to what I did during my PhD study, which was a nice thing to have in a job after graduation.

Model development roles are also available sometimes at banks, which are as the title sounds–building models, but these roles I found are a bit rarer and require more coding/software engineering experience.

So, give it a try, and let me know what you think or if you have questions!

Ryoo, Chanyang CERT 202205 APAM (The certification letter saying I got my PhD in applied mathematics)

GSAS_THESIS_ryoo (My PhD Thesis)

 

All glory be to Christ our King! Many thanks to professors, family, and friends!

 

I learned how to use Flask (web development tool in Python), Bokeh (interactive plotting in Python) and Heroku (webpage deployer) to make an interactive web app here: http://jrtdi.herokuapp.com

1. https://arxiv.org/abs/2112.09696
2. https://arxiv.org/abs/2112.09566

Even though arXiv isn’t publication, these are my first arXiv papers that go through some moderation. We hope to publish them through journal soon.

Recently I received a verification certificate from a machine learning course I took from edX:

HarvardX PH125.8x Certificate _ edX

I used R to implement the algorithms. I learned different regression models, classification models, knn, random forest, ensemble methods, k-fold cross validation, and about working and manipulating datasets..

This is a website that displays what I find wonderful, both academically and non-academically.

Feel free to browse, read, and discuss with me.