KeMeT Tech
Field notes

One issue a month. Real engineering notes.

A short monthly note from a working multi-cloud and AI engineering practice. KQL detections that earned their keep. Bicep landing-zone implementation considerations. Marketplace submission guidance. Agent evaluation patterns that survived contact with real users.

Written for senior engineers, not for buyers. Reply to any issue and a senior engineer reads it.

Subscribe

Inbox-first. No tracking pixels. No share buttons.

The newsletter is delivered plain-text-first. No marketing automation stack underneath. No suggested-content footer. The unsubscribe link is in every issue and works on the first click.

One issue a month. No tracking pixels. Unsubscribe link in every issue.

Upcoming issues

The first four issues, queued.

  • 001 · June 2026

    Six KQL rules every Sentinel rollout should ship with

    Walking through the open-source detection pack: what each rule catches, the noise patterns to expect, and which SOAR action belongs on each.

  • 002 · July 2026

    The Azure landing-zone implementation considerations that bounce a HIPAA audit

    Six policy-as-code gaps that pass cfn-nag and tfsec but fail an actual HIPAA assessment. With the Bicep snippets to close each.

  • 003 · August 2026

    AWS Marketplace SaaS contracts: the five cert-rejection patterns

    Why most first-time submissions bounce. IAM-role granularity, telemetry endpoint shape, EULA conflicts, CloudFormation parameter validation. Patterns for preparing and validating a marketplace submission.

  • 004 · September 2026

    Cost routing a multi-model agent: what to ship, what to skip

    How model routing and caching can improve cost efficiency. The router classifier, the prompt-caching layer, the eval harness that catches model-version regressions.