A New Start

It’s been a while.

The world of software development and data science has changed a lot since I last wrote here. AI and LLMs are more commonplace. uv has changed how I manage Python and Python projects and what was considered a "preferred practice" a year or two ago might feel a bit dated today. With all the changes, I felt it was time to give this website a refresh and relaunch.

The goal remains the same: to collect and share thoughts and ideas on modernising the actuarial toolkit. I’m excited to get back to it.

Updating the "Python Preferred Practices"

The first step was to revisit the Python Preferred Practices (PPP) series. uv has had a massive impact on the Python ecosystem, and I use it extensively now. So, I’ve updated the PPP posts to reflect this new reality.

The most significant change is the move to uv for both Python version management and dependency management. If you've read the original posts, you'll remember the workflow involved a combination of pyenv, virtualenv, and pip-tools. While that approach is still perfectly valid, uv has simplified things immensely by rolling all that functionality (and more) into a single tool.

• PPP 1: Python Version Management has been updated to reflect how uv can install and manage different Python versions.
• PPP 2: Dependency Management now demonstrates how to use uv with a pyproject.toml file to create virtual environments and manage project dependencies in a clean, modern way.

You might also notice that PPP 3: Project Automation has been deprecated. With uv handling so much of the environment setup automatically, the old Makefile approach felt a bit heavy for simple projects. Automation is still an important topic for more complex workflows, and I plan to revisit it in a future post.

Archiving and Moving Forward

To keep the main page focused and relevant, I’ve moved many of my older posts to an archive. I also removed the apps I used to host. They were fun to make and served their purpose as a learning exercise for me. I tried to rebuild them for the new site, but they were taking too much time, and I figured it was better to focus on writing about more relevant stuff.

Looking Ahead

My intention is to contribute more regularly from now on. I have evolved greatly since the last time I posted. My current role, where I work on climate modelling for Old Mutual Insure, has taken a lot of my focus. However, it developed my skills greatly. I have also written three longer form papers since posting here, two on LLMs, and one on mortality modelling.

I think AI is a double-edged sword. It is immensely useful and beneficial if used correctly, it can also be a huge hindrance to learning, quality, and uniqueness if used poorly. I plan to write more about this in future posts.

Thanks for reading, and stay tuned.