2025-CAS-RPM-NN-GenAI-Workshop

Setup

1. Clone the Repository:

git clone https://github.com/bettyzhu912/2025-CAS-RPM-NN-GenAI-Workshop.git

2. Check you have Python installed:

Make sure you have Python >= 3.9 installed. If you're on a work computer, you may need admin privileges to install Python.

You can check you have python by running:

python --version

3. Install the required packages:

Mac/Linux:

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Windows:

python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt

4. Run the notebook:

jupyter notebook

Then you can open the notebook in your browser.

5. API Keys:

This notebook uses an OpenAI-based approach for simplicity. Given the low token usage, it should only cost a few cents to run.

Shared API Key:

We will share a token with you to use for duration of the workshop.

Personal API Key:

If you'd like to rerun this notebook on your own, follow these steps:

• If you're new to the OpenAI API, sign up for an account.
• Follow the Quickstart to retrieve your API key.

You can paste that key into the notebook in the cell that starts with

os.environ["OPENAI_API_KEY"] = "YOUR_API_KEY"

Make sure to replace YOUR_API_KEY with your actual API key and uncomment the line (remove the # at the beginning of the line).

⚠️ Work Computer SSL Issues: Some corporate networks may block access to OpenAI’s API, leading to SSL verification errors. If you encounter this issue, consider:

• Running the notebook on a personal computer.
• Using the hosted Colab version (ensure your VPN is off).
• Checking with your IT team for potential workarounds, such as configuring SSL settings or using a company-approved proxy.

Part 1: ...

Part 2: Bridging Data Divides Workbook

This repo offers a hands-on workbook and interactive notebook examples that illustrate key concepts such as retrieval-augmented generation, embeddings, and structured outputs. It’s tailored for actuaries and data professionals seeking practical approaches to unify and analyze unstructured data, apply AI techniques to real-world challenges, and integrate these insights into existing insurance workflows.

Hosted Colab Version:

For those who don't want to setup Python, you can use the hosted Colab version of the notebook. Note this does require a Google account and it may be blocked by your company's security policies.

Open In Colab