Blog

An evolution of the Terraform data structures pattern — separating WHAT from HOW from WHERE for smaller teams and AI-assisted development.

How migrating from tightly-coupled Bedrock API calls to a pydantic-ai Agent decouples the AI integration from platform-specific code — making provider switching trivial, silent failures explicit, and tool integration possible.

How to integrate vision inference into serverless architectures without specialised ML infrastructure using AWS Bedrock, Pydantic validation, and Parquet storage.

How I used the Repository pattern to decouple a hand-built monitoring platform from its database backend — and what that made possible when it came time to migrate.

How stacking non-linear layers solves XOR, what loss and gradient descent actually do, and a from-scratch NumPy implementation of backpropagation.

Why activation functions exist, how the common ones differ, and which I'd reach for — with a Python demo showing the effect of swapping activation functions on a small network.

An overview of the Perceptron and its influence on modern neural networks.

A write-up of my thoughts of CI/CD at scale as presented at various talks and conferences.

How applying data structures and algorithms to Terraform turns infrastructure from a bottleneck into a self-service platform — with working code and a real proof point.

How I replaced a Docker/EC2/MySQL monolith with a fully serverless REST API — and made the Python as data-driven as the Terraform.

A little primer on said pattern, with some quite silly examples.

One OpenAPI spec driving both the Terraform that provisions an API Gateway and a SwaggerUI documentation site — with CORS as the main gotcha.

How I used Terraform to manage all of Sudoblark's GitHub repositories — covering the three-tier data pattern, bootstrapping paradox, and drift detection.

In 2023 I bootstrapped Terraform for my own AWS account. Three years on, looking back at that setup is professionally embarrassing — long-lived credentials, a single account, no tests. Here's what I built instead, and what I'd tell someone starting today.

I once tried to make a terminal dungeon crawler. It never shipped. But the monster name generator I built for it became a surprisingly functional REST API — and taught me more about BDD testing than any tutorial ever did.

How a three-parameter recursive function handles pagination across almost any REST API — page numbers, next-link cursors, nested responses, and the cases where it still breaks.