AI and Autonomous Systems Report

Online edition · July 2026
Lead
How I ship AI workflows that survive contact with operations

A field guide to the gap between prototypes and the warehouse floor: n8n, AWS runbooks, and the unglamorous layer that keeps agentic tools honest.

Also
Travel Avenue, from blank repo to a storefront that must sell

Catalog, media, SEO, and the automation underneath a UAE luggage DTC channel.

Research
What markerless motion capture taught me about evaluation honesty

Metrics that win a slide can still lose the lab. Notes from a computer-vision thesis path.

Opinion
The quiet craft of internal tools nobody demos on Twitter

WorkOS, Docker Compose, and backups over another chat UI.

Tips of the day

  • Make every job safe to re-runIdempotency first — an automation you can replay is one you can actually trust.
  • Log the agent's every moveIf a workflow touched production, you should be able to replay exactly why.
  • Test the restore, not the backupA backup you have never restored is a hope, not a plan.
About

I write these between shipping things — notes on the AI and autonomous systems I actually build. If a post sparks something, the projects and research behind it live on my portfolio.

See the portfolio →
More stories
NLP
MultiSent and the limits of “just fine-tune it”

When multilingual sentiment needs more than a single checkpoint and a hope.

Career
From research lab to Dubai operations: what transferred

A short map of skills that survived the jump into automation engineering.

ML
Diabetes risk models: ROC curves and clinical restraint

What a clean dashboard can hide about thresholds and leakage.

Collections