The growth of AI and LLMs in Python has given the language a whole new life. It’s been my favourite language for years and it always makes me happy to see it grow (and that’s despite whatever Theprimeagen and DHH might say!). I came to Python through Django — moving on from systems programming in C++. Django clicked instantly due to fast POCs, smooth dev experience, and the nature of interpreted languages (And gosh, the .h files. I don’t want to deal with .h files any more!)
As I see it, Python’s usage broadly comes down to two major applications:
- Web backends and
- data processing, AI, ML, LLMs
For a long time, I had a very elementary understanding of Python’s use in data science. It wasn’t until I got hands-on with data engineering and ML that I really saw what that side of the ecosystem looked like.
One thing that stands out to me is that the two worlds use the same language, but in different ways. Python code in backend projects tends to be more structured. I’ve seen better scaffolding, use of MVC, stronger emphasis on OOP, and stricter typing. Production backends bring a different architectural discipline because deployment and uptime force it.
In contrast, the data/ML side often starts from notebooks, ad hoc scripts, and pandas hacks. Great for exploration, but it can come at the cost of maintainability.
It’s fun to bring that backend mindset into the mix on data engineering projects. Knowing how systems are deployed, how services are structured, and how things run in production (Kubernetes/AWS et al) makes it easier to think ahead about the workflows. It feels good to have finally connected the dots between the backend and data world. I feel there’s a lot more value in that overlap than most people realize. Stay tuned for some upcoming posts and projects if you’re excited about Python, AI, and Data Science.
Leave a Reply