Open Brain
AI EngineerVideoSource Linked

Agents Made Code Free. Maintaining It Didn't Get Cheaper. - Dmitry Buykin, Maersk

When agents write most of a production codebase, the bottleneck shifts to review and maintenance. Buykin describes weekly evidence loops for keeping changes understandable and safe to ship.

AI Engineer · Nov 11, 2026
Open Source Open MarkdownOpen JSON
Source Summary

Buykin’s system runs across **110 countries** and about **10,000 scenarios**. It keeps expert-owned rules readable, gives agents only task-relevant context, and uses runtime checks to constrain changes. A replay of a proposed check against thousands of past sessions would have triggered on less than **1%**.

Practical Implication

For coding-agent workflows, make repeated agent paths explicit tools, preserve human judgment where rules vary, and connect user feedback to evaluations. Keep durable memory in files and require evidence such as replaying real tasks; a second agent can check changes against a checklist without seeing the author’s reasoning.

Agent-Ready Context
Buykin’s system runs across **110 countries** and about **10,000 scenarios**. It keeps expert-owned rules readable, gives agents only task-relevant context, and uses runtime checks to constrain changes. A replay of a proposed check against thousands of past sessions would have triggered on less than **1%**.

For coding-agent workflows, make repeated agent paths explicit tools, preserve human judgment where rules vary, and connect user feedback to evaluations. Keep durable memory in files and require evidence such as replaying real tasks; a second agent can check changes against a checklist without seeing the author’s reasoning.

The approach depends on people reviewing the decisions agents cannot reliably infer, and on having real tasks to replay. The talk describes one production system; it does not establish that the same checks or thresholds fit every codebase.
Connected Context · Feed7 Judgment

This account emphasizes how production reliability depends on keeping expert rules readable, limiting agent context, and replaying proposed checks against historical sessions. The reported under-1% trigger rate illustrates how replay can narrow a check before deployment, without establishing that the threshold suits other systems.

Tribal Dungeons of Global Shipping: AI Agents at Global Scale — Dmitry Buykin, MaerskBoth describe Maersk production agents and emphasize executable, bounded procedures plus expert correction; this signal adds readable rule ownership, task-scoped context, and replay of checks.Catching Bugs Before They Land: JiTTesting’s Human-on-the-Loop Approach at Full ScaleBoth use historical or behavioral evidence to filter checks before code changes proceed, with human judgment retained; JiTTesting describes risk-ranked generated tests, while this signal highlights replaying a proposed check against past sessions.Procedural Graphs: Self-Evolving Execution Structures for LLM AgentsBoth make repeated agent work more explicit and reviewable; this signal stresses expert-owned readable rules and replay, while Procedural Graphs proposes evolving executable workflows and calls for held-out validation of updates.
Context Map
agentcoding#harness-engineering#agent-reliability#agent-memory
Uncertainty
The approach depends on people reviewing the decisions agents cannot reliably infer, and on having real tasks to replay. The talk describes one production system; it does not establish that the same checks or thresholds fit every codebase.