AI
Engineering & Architecture Notes
Production AI,
Platforms & Systems
Practical patterns for building governed AI agent harnesses, scalable platform architectures, homelab infrastructure, and developer workspace setups. Distilled from real-world engineering experience.
By Shashi Kanth G S, Software Architect, Airline domain, Enterprise AI, Platform EngineeringRecent Blogs
View all
Explore by Category
Curated technical domains and deep dives
AI & Agent Governance
Harness boundaries, privilege brokers, sandboxing, memory isolation, and production evaluation loops.
21 Articles
Platform Architecture
Distributed systems design, event-driven pipelines, failure containment, and scaling trade-offs.
6+ Articles
Home Lab & Infra
Self-hosting, bare-metal servers, Kubernetes clusters, Proxmox, ZFS storage, and home datacenter networking.
Active Category
Desk Setup & Gear
Developer workstation ergonomics, multi-monitor layouts, mechanical keyboards, and tech gear.
Active Category
More Articles
AI
July 18, 2026
6 min read
Output Scoring Passes The Agent That Leaked Your Credentials
A correct answer reached by a forbidden path is a failure. Assert on the trajectory and the policy decisions, not the final message. Runnable suite.
AI
July 14, 2026
6 min read
Your Tool Definitions Are Prompts, And Most Are Bad Ones
An unbounded tool response costs 17x more tokens for the same answer. A runnable auditor for the four tool-design mistakes that make agents guess.
AI
July 11, 2026
6 min read
"I've Fixed The Failing Test" Is A Claim, Not A Completion Signal
Self-reported completion is not evidence. Independent deterministic verifiers, feedback re-injection, and why an LLM judging itself is theatre.
AI
July 7, 2026
6 min read
The Retry That Paid The Invoice Twice
At-least-once execution is the default everywhere. Checkpoints make resumption cheap; idempotency keys make it safe. Runnable code and the bug I hit.
AI
July 4, 2026
6 min read
Context Rot: When Your Agent Forgets It Was Denied
Dropping the oldest messages is amnesia with a token budget. Compact by turn group, pin the goal and the denials, and never orphan a tool call.
AI
June 30, 2026
6 min read
A Regex Is Not A Sandbox
Isolation comes from containers. What belongs in your agent's design is the policy layer: resolved-path containment, allowlists, network off by default.