Chip Huyen

Things Chip Says on Podcasts

I work at the intersection of AI, education, and storytelling. I help companies deploy AI applications in production. I'm the author of Designing Machine Learning Systems and AI Engineering, but I also write novels and stuff.

Where to Find Them

Chip Huyen writes The Pragmatic Engineer and writes Chip’s Substack . They have also been a guest on Super Data Science: ML & AI Podcast with Jon Krohn (2 times) , Gradient Dissent: Conversations on AI , The Gradient , The InfoQ Podcast , Lenny's Podcast: Product | Career | Growth , The MAD Podcast with Matt Turck and The TWIML AI Podcast .

Recently: “AI, ML, and Data Engineering InfoQ Trends Report 2026” on The InfoQ Podcast (October 2026); “999: What's Left to Build When Software Is Free, with Chip Huyen” on Super Data Science: ML & AI Podcast with Jon Krohn (June 2026); “Al Engineering 101 with Chip Huyen (Nvidia, Stanford, Netflix)” on Lenny's Podcast: Product | Career | Growth (October 2025); “The AI Engineering Stack” on The Pragmatic Engineer (May 2025); “AI Engineering with Chip Huyen” on The Pragmatic Engineer (February 2025); “AI Engineering Pitfalls with Chip Huyen - #715” on The TWIML AI Podcast (January 2025).

What They Said

During a panel discussion of multi-agent systems, she described her own experiment in getting several coding agents to coordinate on one task. She then asked the panel how to make an agent recognize that a task is its own to do.

The InfoQ Podcast · 2026-10-30 Permalink → Listen →
The InfoQ Podcast Around 52:38 into the episode
Speaker 3

…like is this like one orchestrator that's spin up like multiple workers? Or like, do you have like different agents working in parallel? And then how do you, and if that's the case, and how do you deal with communications like across different agents working on the same kind of like complex task?

Speaker 4

…and they have to do that in a social manner, which means communicating, so FIPA standards and things like that, everything kind of revolves around orchestration communication. So I have different protocols for communication, different architectures for orchestration, a lot of work on just this too.

Speaker 3

Yeah, so I'm very interested in like agent to agent communications, right? So actually recently, I use multi-agent a lot. So I have some complex tasks. I would like get a bunch of agents to work together. So I actually built this protocol to let them like, okay, when you start working on a task, it's a part of a big system. So like other agents working on the same task, so we can message them to make sure that you guys go, like you can coordinate and like don't overlap. So what happened after that is that like my…

Speaker 4

You try, you don't do it. You know, you try to do that. Because the thing is, depending on how you're tackling problems, for example, one of the simplest ways of doing that is using a blackboard kind of messaging to see what everyone's doing. So they're just posting backs and I have the shared…

Speaker 3

No, it's a very hard problem. I'm wondering, like, if any of you, like, I'm supposed to hear from here and Daniel, like, have you used the box orchestrations ability of different models? First of all, like for in cloud or in codecs where you can ask an agent, okay, spin up my workers to do this…

Speaker 1

I haven't done that extensively. Maybe, Daniel.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

Collections They Appear In