Insights & Essays.
Field notes for CTOs and principal engineers: what shipped, what failed, and how to staff the next AI program. Depth first — no vendor recap.
Latest Articles
OWASP Top 10 for Agentic AI (ASI01-ASI10): Implementation Guide for Enterprise Teams
Master OWASP's Top 10 for Agentic AI (ASI01-ASI10). Learn goal hijacking, tool misuse, runtime governance, MCP verification, Lakera Guard, and Garak red-teaming.
Cursor Cloud Agents, /multitask, and Parallel Worktrees: Running 8 Agents Simultaneously in 2026
Master Cursor Cloud Agents, /multitask, and Git worktrees. Learn to run 8 parallel AI agents in cloud VMs without file conflicts or token bloat.
Mastering CLAUDE.md, AGENTS.md, and SKILL.md: Anthropic's 7-Method Steering System for Claude Code
Learn Anthropic's definitive 7-method steering system for Claude Code. Master CLAUDE.md, AGENTS.md, SKILL.md scopes, token costs, specialist pool, and gatekeeper patterns.
The 70/30 Rule: How High-Performing Teams Actually Do AI-Assisted Development
The 70/30 rule for AI-assisted development: a practical framework for engineering leaders to govern what AI executes autonomously versus what humans must own. Architecture, metrics, governance patterns, and the 2027-2030 roadmap.
Managed vs Self-Hosted Agent Harness — The Deployment Trade-off Engineering Leaders Get Wrong
Managed agents ship Tuesday. Self-hosted agents ship architecture. Pick wrong and portability becomes a fantasy slide. Here is how to evaluate the TCO, compliance, and coupling.
Platform Engineering for the Agentic AI Era — When Your IDP Becomes the Governance Machine
78% of enterprises have a platform team. Only 23% have an IDP that can govern agents. That gap is where shadow agents live. Here is how to rebuild your IDP.
DORA Is Not Enough in 2026 — The AI Attribution Layer Engineering Leaders Actually Need
DORA metrics celebrate deployment frequency but mask massive rework rates. Here is the AI Attribution layer engineering leaders need to audit real impact.
Claude Code Hooks Deep Dive: PreToolUse, PostEdit, and PreCompact for Zero-Drift Agentic Codebases
Unlock the full power of Claude Code hooks. Master PreToolUse security interceptors (exit code 2), PostEdit automated linting/testing, and zero-token PreCompact context backups to eliminate AI hallucination drift.
Continuous Discovery Habits in the AI Age: Teresa Torres Framework Updated
Learn how to update Teresa Torres' Continuous Discovery Habits using AI-assisted interview synthesis, automated assumption testing, and opportunity.
Questions
What is this blog for?
Long-form notes on agentic AI, production LLM architecture, cloud engineering, and technical leadership, authored by Vatsal Shah. Each article has its own title, date, and canonical URL.
Are articles free to read?
Yes. Posts on shahvatsal.com/blog are free, with no paywall. Related case studies, playbooks, and frameworks sit in their own hubs.
How do hiring or consulting readers use this hub?
Hiring: read the article then /resume or /contact?focus=hiring. Consulting or software delivery: /contact or /contact?focus=software-application. Magento claims are not marketed here.
External Standards & Publications
Our engineering publications follow industry standards for software architecture and agentic AI systems. Read more about technical specifications at the IEEE Standards Association and general web engineering principles at the World Wide Web Consortium (W3C).