Loading video...

Video Failed to Load

Go Home

Andrew Ng just released a 2-hour course on full Graph Engineering. How to go from one prompt to 100 agents that loop, rewrite themselves, and run without you: 09:14 - Build your first AI agent 33:11 - Run agents with loop engineering 1:02:46 - Turn agent loops into graphs...

227,687 views • 3 months ago •via X (Twitter)

36 Comments

Neuromancer.eth | 1972.eth | satan.base.eth's profile picture
Neuromancer.eth | 1972.eth | satan.base.eth3 months ago

ENS

elune's profile picture
elune3 months ago

the unattended graph section is probably worth the full two hours alone

Lunar's profile picture
Lunar3 months ago

yeah that section is gold

KravX🧩's profile picture
KravX🧩3 months ago

more agents only help if routing and evaluation are strong enough to catch the errors they multiply

Lunar's profile picture
Lunar3 months ago

routing and evals are everything

Valentin's profile picture
Valentin3 months ago

Blyat why he’s Andrew? He’s Ching Lu

Lunar's profile picture
Lunar3 months ago

nope, that's andrew ng

0xSlyth's profile picture
0xSlyth3 months ago

gold

Lunar's profile picture
Lunar3 months ago

thats true

Elizaveta Zabrodskaya's profile picture
Elizaveta Zabrodskaya3 months ago

Watching it tonight

Lunar's profile picture
Lunar3 months ago

perfect for tonight

Mikadzyki🌙's profile picture
Mikadzyki🌙3 months ago

that progression is the roadmap

Lunar's profile picture
Lunar3 months ago

yeah thats the roadmap

Saeed Anwar's profile picture
Saeed Anwar3 months ago

Graph engineering is where most agent projects either compound or break. Did Ng cover state management when graphs get nested more than two layers deep?

Paulz's profile picture
Paulz3 months ago

@grok give me the link of the mentioned course about Graph Engineering by Andrew Ng.

Suede Labs AI's profile picture
Suede Labs AI3 months ago

Loops create motion; graphs create inspectable state. The next maturity step is giving each node an evidence contract: inputs, outputs, evaluator, approval boundary, and artifact trail. Otherwise 100 agents just multiply unverifiable work.

Leo Oliemans | Refinery's profile picture
Leo Oliemans | Refinery3 months ago

A loop that can rewrite itself also needs a way to stop when the world changed underneath it. Do you model freshness or conflicting source state as a first-class exit, or only tool/error failures?

Quant-Builder.ai's profile picture
Quant-Builder.ai3 months ago

He’s the man. Took a bunch of his classes. What a legend

الكتلة_ روان منيرة الثبيتي's profile picture
الكتلة_ روان منيرة الثبيتي3 months ago

Her smile turns grace into a quiet power

Samurayich's profile picture
Samurayich3 months ago

Three pages of notes. The 1:49:05 timestamp answers a question I've been chasing for 2 months: how to make a render pipeline refine itself without manual intervention.

Andrius Mostauskas's profile picture
Andrius Mostauskas3 months ago

It’s probably best suited to content creators. Their workflow involves repeatable patterns. I can’t see using this in my work/life anytime soon

琦玉老师's profile picture
琦玉老师3 months ago

从 prompt → loop → graph 这条路径很符合实际体验。很多人卡在 loop 阶段就上不去了——agent 跑得动但没法 scale,因为 loop 之间的状态传递和错误恢复没人管。graph engineering 本质上是在给 agent 加拓扑约束,比裸跑 loop 可控太多

AM's profile picture
AM3 months ago

He has lost the capability to read something and research and then write essay on it. Who says researching and writing is boring and monotonous task 🤣🤣 Oh i forgot just because he is not good at it.

workflowfix's profile picture
workflowfix3 months ago

This is the actual ladder. Most people stop at “I built an agent.” The real work starts when you move to loops, then graphs, then systems that can improve without you sitting on them. Single agents are the old workflow. Self-improving systems are the new one.

Xingheweiwei's profile picture
Xingheweiwei3 months ago

good

Fidelity's profile picture
Fidelity3 months ago

Agentic workflows != Graph Engineering

Diogo Perillo's profile picture
Diogo Perillo3 months ago

Take a look at how works, you can do a lot of that there

Nara Shin's profile picture
Nara Shin3 months ago

what a time we are in now, writing loops is now an engineering field. there is literally nothing new about this course…

morph's profile picture
morph3 months ago

if you use AI agents and haven't even seen this video guide before I guess you just waste your time

netcinemo's profile picture
netcinemo3 months ago

Let's Loop n Flow!! 🎉✨

AJawhar's profile picture
AJawhar3 months ago

Very informative. But do you think it is like hype more than like a best practice?

Abhishek K Shankar's profile picture
Abhishek K Shankar3 months ago

@grok Can you provide me the link of course referred here?

DeDi's profile picture
DeDi3 months ago

从提示到自改写智能体,两小时课程把图工程讲透了

Saman Ahmed's profile picture
Saman Ahmed3 months ago

I’d spend most of the time on loop engineering.

Adel Bucetta's profile picture
Adel Bucetta3 months ago

the honest answer is most devs don't even know where to start with graph eng, never mind scaling it to 100 agents

AI Apps API's profile picture
AI Apps API3 months ago

Loops and graphs are the right unit to teach, and they are also where the bill quietly forms. A loop that resends the same system prompt and tool definitions every pass is paying full price for tokens that never changed. Prompt caching covers most of that, and batching anything nobody is waiting on covers a good part of the rest. Cheaper to wire in before the graph gets wide than after.

Related Videos