Field notes from the work
Posts
Concise lessons on architecture, developer tooling, cloud tradeoffs, and the work of moving software from intent to adoption.
How I Would Manage Enterprise Copilot Agents: Plugins, Managed Settings, and APM
A practical architecture for packaging, distributing, and governing GitHub Copilot agents across an enterprise.
Make Every Token Count: Why Agent Quality Is the Better Way to Optimize Spend
GitHub Copilot switched from premium requests to usage-based billing. That single change set off a...
Squad vs Fleet Mode in GitHub Copilot CLI: What's the Difference?
TL;DR For the Busy Dev Fleet Mode (/fleet) is built into Copilot CLI. It auto-decomposes...
GitHub Copilot SDK vs Azure AI Foundry Agents: Which One Should Your Company Use?
TL;DR for the Busy Person GitHub Copilot SDK gives you the same agentic core that powers...
Building Context-Aware CI with GitHub Copilot SDK and Microsoft WorkIQ
Overview Building Context-Aware GitHub Actions with the Copilot SDK and Microsoft...
GitHub + Azure DevOps: The Better Together Story (And Why GitHub Should Be Your Future
🧠TL;DR for the Busy Person — Why GitHub Should Be Your Long‑Term SDLC Home (Even If...
What If Your CI Pipeline Could catch regulatory compliance violations of your code?
TL;DR for the Busy Dev: Traditional CI checks (linting, tests, SAST) are...
Injecting AI Agents into CI/CD: Using GitHub Copilot CLI in GitHub Actions for Smart Failures
TL;DR for the Busy Dev We are used to CI/CD pipelines that fail on syntax errors or failed...
Vibe Coding is Technical Debt. Vibe Engineering is the Fix
TL;DR for the Busy Dev Vibe Coding is "Single Player Mode": Prompting based on...
Architecting for Speed & Safety: How We 2x'd Dev Velocity with ts-rest and Next.js API (and supercharged our AI copilot)
The tl;dr for the Busy Dev Our initial backend was technically robust but slow to adapt to...
So… what is GitHub Copilot’s "Raptor mini"and why should devs care?
GitHub quietly slipped a new model into Copilot called “Raptor mini (Preview)”. The changelog...
I'm Going All-In on AI for Developers.
It feels like every developer is drinking from a firehose of AI information right now. We're seeing...
A CTO’s View on AI Regulation: Canada’s Second Chance for a Competitive Approach
A founder and CTO’s case for interoperable Canadian AI regulation that protects safety without duplicating the compliance burden on startups.
How to Save $17,000 a Month by Self-Hosting DeepSeek R1 on Google Cloud Run GPU
A February 2025 walkthrough of self-hosting a distilled DeepSeek R1 model on Google Cloud Run GPU for spiky, pay-per-use inference workloads.
Cloud Cost Cutting Myths Busted: How We Unlocked Huge Savings With Minimal Effort on GCP
A 2024 Adauris case study on low-effort changes to logging, storage, Cloud Run, Compute Engine, and legacy data hosting.
Don’t Let Legacy Data Break the Bank: How we reduced a data hosting cost from $4080 annually to $1.8 per year
How Adauris preserved legacy MongoDB engagement data by moving a cluster snapshot to S3 Glacier and cutting recurring storage costs.