Sam and ExamPA.net
Hi! I’m Sam, an actuarial predictive modeler.
You probably found ExamPA.net because you’ve worked through the SOA’s PA modules and asked yourself “isn’t there a faster way of studying?”
What a coincidence! Three years ago, I said this same thing to my friend who was also studying for PA. It was three weeks before PA, and we were cramming. We were both good students who had good study habits, but our consensus was that there were not enough practice problems to work on.
- Both of us had been through the modules twice, and still had lots of unanswered questions.
- We had done all of the practice exams but were still finding topics that had never been tested before.
- Both of us wanted to put this exam behind us quickly. I was 25 years old and my friend was 28, and we had both worked through 5 exams to get to this point.
- We had searched Actuarial Outpost and Reddit and no one had the answers.
We teamed up by combining our study notes and quizzing each other. A few weeks later, we both passed with flying colors. Then only a few weeks later, two coworkers asked me if I had any notes and so I posted them on LinkedIn. Then dozens of people began to message me asking for them. It was like being in a college class with an impossible professor which causes the students to form a study group and share notes. At this point I decided to create an online course.
Data science tools, documentation, and training resources (Kaggle, Coursera, Udemy, AI for everyone, etc.) have advanced rapidly over the last few years, but there is still a high demand for these resources within actuarial science.
The way that data scientists learn machine learning, and how I learned at first, was not by taking an exam but by doing lots and lots of examples: kaggle competitions, data science interview projects, reading textbooks, and watching tutorial videos. This is a different style than the usual actuarial study method of buying a single study manual and then reading it from front-to-back. The goal of exampa.net is to give everyone access to this style of learning.
Sam has worked in predictive modeling for four years. At Milliman Intelliscript, he developed machine learning-based risk scores for health insurance. At Willis Towers Watson, he did modeling for customer loyalty programs such as airlines and hotel chains. These clients had “big data” on the scale of hundreds of terabytes. All of the “free airline miles” from customers all over the world added up to liabilities in the billions of dollars. He now does actuarial consulting in life and health modeling while being an instructor for PA.
Read my first place submission to the SOA's Predictive Analytics Contest
“The Predictive Analytics and Futurism and Technology Sections invite you to participate in the inaugural “2019 Predictive Modeling, Innovation and Industry Contest.” Analytics are the future of insurance and this contest will recognize the best and brightest who can apply their skills to design an analytical solution that solves an insurance problem, submitted in a Jupyter Notebook.”