Another Epoch

Since ChatGPT dropped in late 2022, AI progressed in lightning speed. Today, generating running code has never been easier - with the help of AI assistant, you can create a data pipeline, train or connect to a model in a fraction of time it used to take. The technical barriers are disappearing fast.

Nowadays, the real challenge isn’t whether you can do it — it’s how you do it. The choices you make in system design, data handling, training pipelines, and architecture determine whether your AI and data systems are reliable, scalable, and trustworthy, or fragile, biased, and prone to failure.

How do you architect systems that support reliable AI or data products in production? What biases lurk in our models, and can system design choices help control or mitigate them? How do you build data pipelines that scale easily while staying robust and secure, continuously transforming data to feed everything from dashboards to training models? What’s the theory behind it, and what are the trade-offs?

These are some of the questions I will explore in this newsletter. If it sounds like something you would enjoy, subscribe :)
Let’s talk about the boring interesting stuff! ^^

About me

Hi, my name is Sarah and thanks for stopping by!

In my spare time, I’m usually moving — I love all kinds of sports, and right now, I’m especially into climbing. I also enjoy cooking and experimenting with new ingredient combinations (often making a total mess in the kitchen, but that’s part of the fun, I guess ^^).

As you’ve probably guessed from the fact that I created this newsletter, I’m really interested in AI, data engineering and everything in between. Another Epoch is my way of sharing what I’ve learned while continuing to explore and deepen my understanding along the way.