Python on Home

Speaking at Pycon UK 2025

Wed, 01 Oct 2025 00:00:00 +0000

Last month, I had the opportunity to speak at PyCon UK for the very first time—and what an experience it was!

For those who haven’t attended before, PyCon UK is more than just a conference about Python. It’s a celebration of community. Each year, it brings together people from diverse backgrounds—engineers, researchers, educators, hobbyists—to share ideas, learn from one another, and showcase the many ways Python is being used across fields.

Python: The Documentary

Thu, 04 Sep 2025 00:00:00 +0000

I just watched “Python: The Documentary” and absolutely loved it 🩵 Highly recommended!

Takeaway

The documentary narrates not only the technical milestones but the human journey:

🚀 How Guido van Rossum’s vision and persistence turned a small project (started over a Christmas break!) into a global phenomenon.
🚨How the community navigated tough transitions (obviously we are talking about Python 2 → 3) and came out stronger.
🐍 How tools like Anaconda helped Python explode in data science.
✨ And how a culture of inclusivity transformed PyCon into a space that reflects the diversity of the wider world.

It’s fascinating to see that Python’s growth wasn’t just about clever design choices. Its real strength came from something deeper: A culture that welcomes ideas, contributions, and people. As Brett Cannon famously said:

When GPUs Make Python Slower

Fri, 22 Aug 2025 00:00:00 +0000

Just finished watching this interesting talk from the PyconUS 2025 Conference.

Takeaways

Minimise memory transfers:

The GPU is a separate device with its own memory (most modern GPUs rely on VRAM). To process data, you often need to transfer it from CPU (host) memory to GPU (device) memory and back again. These transfers happen over the PCIe bus, which is much slower than the GPU’s internal memory bandwidth.
Pitfall: If you repeatedly move data back and forth in small chunks, the transfer latency dominates, and the GPU spends more time waiting than computing. This can make your code slower than a pure CPU implementation.

Maximise operational intensity:

aleatory

Sun, 01 Jan 2023 00:00:00 +0000

Overview

Aleatory (/ˈeɪliətəri/) introduces objects representing stochastic processes, and provides functionality to:

fanchart

Sun, 05 Jan 2020 00:00:00 +0000

Overview

The term fanchart was coined by the Bank of England in 1996. Since then, central banks around the world have used these charts to illustrate its forecasts for inflation, GDP, and other economy measures.

two-piece

Fri, 11 Jan 2019 00:00:00 +0000

Overview

The twopiece library provides a Python implementation of the family of Two Piece distributions. It covers three subfamilies Two-Piece Scale, Two-Piece Shape, and Double Two-Piece. The following diagram shows how these families relate.