Founder · Building AI-Native Organization

Hi, I'm Limin Ge. Founder & Engineer building products people use, play, and pay for.

I work on agentic systems, context engineering, large-scale distributed systems, and building AI-native organizations.

Featured articles

A starting list of essays and notes.

View all articles

Business & OrganizationJul 28, 2026

Four Mornings of a Retainer Team: The Four Levels of Enterprise AI-Native Maturity

AI-native is not a scale of how much of the old org you have automated — it is a redesign of how people, agents, information, and decisions relate. From four mornings to why a digital employee can charge $20K–200K a year.

Engineering & ProductivityMay 25, 2026

SSOT: Why Every Software Engineering Principle Is Really About One Thing

Complex products, AI-native output, and fast iteration naturally split team truth. The SSOT Nine-Layer Pyramid lets AI agents, product, operations, sales, support, and engineering share one current decision snapshot.

ManifestoDec 4, 2025

AI-Native Engineering: 11 Hard Truths for 2025

A founder note on building AI-native organizations and systems.

Founder notes

AI-Native Engineering: 11 Hard Truths for 2025

Principles I'm using to build AI-native organizations and products.

Read the manifesto

01If a knowledge, methodology, or tech stack is older than 3 years, it is obsolete until proven otherwise.
02Your codebase, docs, and meeting notes are not your backbone; your test case library is the only compounding asset.
03AI is completely, fundamentally stateless.
04We can accept AI mistakes as long as they do not kill the team before the next model upgrade.
05Software complexity must flatten: from deep vertical stacks to wide horizontal systems.
06The context window is the most critical computational resource every engineer must master.
07Plan-Act, Test-Code, and Doc-Code-Doc are the new working loops of engineering.
08The future of code is not abstraction but tiny, isolated, AI-readable units that stand on their own.
09AI will never solve the first mile or the last mile; those remain stubbornly human.
10AI-generated artifacts are not side effects; they are a new software modality and part of your engineering assets.
11The real power of AI IDEs and agents is not generation; it is ruthless, intelligent context selection.

Current state of LLM Risks and AI Guardrails

Published in arXiv.org on June 16, 2024 · 131 citations on Google Scholar

A survey of the main risks in deploying LLM systems and the guardrail patterns used to mitigate them, including bias and toxicity checks, prompt and agent safeguards, privacy concerns, and the tradeoffs required to ship these systems responsibly.

Read the paper