#ai-engineering

Agentic Engineering Workflow: what I took from the workshop

Two days of working with agents: from context and specifications to vertical slices, tests, and verification.

A separate completion of Workshop - Agentic Engineering Workflows in October 2026: its program, dates, and SULICOM certificate.

I completed SULICOM's hands-on workshop on the systems around AI agents: tools, context, rules, hooks, and verification.

This edition focuses on AI’s impact on product and platform engineering, alongside changes in security and CI/CD.

A practical path from a controlled RSS fixture through structured AI analysis and human approval to atomic bilingual MDX publication.

How nine production incidents became contracts, validation rules, tests, and new workflows in a local-first Engineering Radar.

How approval hashes, UK/EN parity, and atomic rollback protect a bilingual Engineering Radar Digest from unapproved changes and partial export.

How schema-bound AI analysis, deterministic Critical controls, and human review turn a model into a controlled analytical engine.

Why we chose Obsidian, file contracts, and human-in-the-loop controls over an autonomous AI publisher for technical digests.

How source contracts, stable identity, provenance, and bounded recovery turn an RSS stream into controlled Signals in Obsidian.