For practising software engineers

Turn experiments into dependable engineering.

I build systems, test how they behave, and share what I learn so you can apply it in your own work.

Find your starting point
Explore articles

From the working notes

Build it. Understand it. Explain it.

What did I learn?
What evidence supports it?
Where can someone else apply it?

How I approach the work

Start with a question

01 / Experiment

What should I test before committing?

Explore an idea, question the assumptions, and find evidence that helps you decide.

Explore experimentation
02 / Build

How do I build a useful, efficient system?

Architecture, implementation, performance, and the trade-offs behind a working solution.

Explore building
03 / Improve

How do I make the practice dependable?

Testing, delivery, technical leadership, and making hard-won learning reusable.

Explore improvement

Latest articles

A layered ink-and-paper abstraction in indigo, teal and copper. Scattered fragments gather into a richer connected pattern, suggesting individual memories becoming shared knowledge.
Your Agent’s Memory Needs a Promotion PolicyAI for Engineers

Your Agent’s Memory Needs a Promotion Policy

AI agent memory can preserve mistakes as well as lessons. Separate run evidence from reviewed guidance, and decide when a simpler runbook is enough.
2026-10-12 Read More
Terracotta, violet, cream and ochre geometric sculptures float above a flowing cream ribbon against a softly lit neutral background, suggesting distinct business rules supported by shared processing.
Separating Business Rules from Shared Processing LogicSoftware Architecture

Separating Business Rules from Shared Processing Logic

Learn policy and mechanism separation: keep business rules domain-owned while reusing processing logic, illustrated with a tested .NET example.
2026-10-09 Read More
How I run multiple engineering workstreams in the GitHub Copilot app
How I Run Multiple Engineering Workstreams in the GitHub Copilot AppAI for Engineers

How I Run Multiple Engineering Workstreams in the GitHub Copilot App

Run engineering work in the GitHub Copilot app with clear scope, suitable workspaces, evidence checks, and decisions that stay human-led.
2026-09-11 Read More
Browse all articles

Built to be useful

A practical problem. Reusable code.

Projects turn the learning into something you can try. Explore tools, libraries and examples you can use, adapt and learn from.

Explore the projects
Azure OpenAI Image Token Calculator
Azure OpenAI Image Token Calculator

Azure OpenAI Image Token Calculator

Estimate image token usage and processing costs across Azure OpenAI models, compare configurations, and explore how image dimensions and request volumes affect projected spend.
Read More

On my workbench right now

Putting AI to practical use.

My current experiments involve prompting, agents, orchestration, and evaluating whether their results are fit for purpose.

Explore the current AI work
Your Agent’s Memory Needs a Promotion Policy

Your Agent’s Memory Needs a Promotion Policy

Read More
How I Run Multiple Engineering Workstreams in the GitHub Copilot App

How I Run Multiple Engineering Workstreams in the GitHub Copilot App

Read More

The engineer behind the explanations

Hi, I’m James.

I’m a senior software engineer at Microsoft. Hands-on work gives me something substantive to share. Explaining it exposes gaps and sharpens my understanding.

More about how I work

Keep learning with me.

Read the explanations, explore the code, and follow along on YouTube.

Read the articles
Explore GitHub
Watch on YouTube