Agentic Engineering, Clearly Explained in 60 Min by an Ex-Meta L8 Engineer | Kun Chen

2026-06-07 · Source: Behind the Craft · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Emerging Technologies & Innovation · Depth: Intermediate, extended

Summary

Ex-Meta L8 principal engineer Kun Chen has developed a highly efficient agentic engineering workflow, enabling him to ship 20-40 pull requests daily with minimal manual code review. His system, detailed in a recent episode, centers on a "Plan, Code, Validate" paradigm, where AI agents handle most implementation and initial validation. Chen utilizes three open-source tools he built: Lavish, a visual HTML editor for detailed, interactive planning; Treehouse, for managing parallel agent sessions and worktrees efficiently; and No Mistakes, an AI-powered code review tool that scrutinizes agent-generated code in a fresh context window, auto-fixing minor issues and escalating complex problems. This approach shifts human effort to meticulous planning and high-level validation, allowing for significant parallelization of agent tasks and increased development velocity.

Key takeaway

For AI Engineers aiming to significantly scale your development output and maintain code quality, you must re-evaluate your workflow by prioritizing meticulous planning and robust, automated validation. Delegate coding entirely to agents, focusing your efforts on crafting detailed, interactive specifications and setting up comprehensive AI-driven review pipelines. This approach, supported by tools for parallel agent management and fresh-context code scrutiny, allows you to manage dozens of daily PRs, transforming your role into an orchestrator of high-velocity development.

Key insights

Agentic engineering prioritizes detailed planning and robust validation to scale development velocity with AI.

Principles

• Detailed planning extends agent autonomous work time.
• Parallelize agent sessions to optimize human oversight.
• Use fresh context for unbiased AI code review.

Method

The workflow follows Plan (visual HTML with Lavish), Code (AI agents), and Validate (AI agents with No Mistakes for review, testing, and documentation, escalating product-implication issues).

In practice

• Implement visual HTML editors like Lavish for planning.
• Utilize worktree managers such as Treehouse for concurrent agent tasks.
• Integrate AI-driven code review and end-to-end testing.

Topics

• Agentic Engineering
• AI Agents
• Automated Code Review
• Software Development Workflow
• Parallel Agent Execution
• Visual Planning
• Worktree Management

Code references

• kunchenguid/lavish-axi
• kunchenguid/treehouse
• kunchenguid/no-mistakes

Best for: AI Architect, CTO, VP of Engineering/Data, AI Engineer, Machine Learning Engineer, Software Engineer

Related on AIssential

• "My agentic engineering workflow" | George Pickett (MTS @ Parallel) — Clarifying intent and creating durable artifacts are crucial for effective, long-running AI agent workflows.
• Getting Humans Out of the Way: How to Work with Teams of Agents — Effective coding agent management requires designing autonomous, parallelized, and self-validating systems, not micromanagement.
• AIE Europe Day 2: ft Google Deepmind, Anthropic, Cursor, Factory, Linear, HF, Cerebras & more — AI engineering is rapidly evolving, demanding new approaches to agent orchestration, code quality, and human-AI collaboration.
• Make All Your Tokens And Your Brainwork Count — Iterative agentic coding, driven by human-refined specs, accelerates 0-to-1 prototyping by valuing human input over cheap AI token generation.
• OpenClaw Creator & Agentic Engineering — Optimizing human-agent interaction shifts focus from token/CPU limits to managing human attention and verification loops.
• A Year Of Vibes — Agentic coding tools are fundamentally reshaping software engineering workflows, challenging traditional practices and human-machine relationships.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Behind the Craft.