The Road to Self-Driving Networks with Marvis

In this episode of The Q&AI Podcast, Bob sits down with Sudheer Matta, Senior VP of Product Management at Juniper Networks, to explore the architectural evolution of AI-native networking. They unpack the journey of Marvis®—from anomaly detection to autonomous remediation—and what the Marvis Actions 2.0 facelift means for network design.   

Sudheer shares how Mist™ is building trust in AI by giving users visibility and control, and how this shift is laying the groundwork for fully self-driving networks. Tune in as Bob and Sudheer discuss the strategic implications of AI in network architecture and what it takes to design networks that can think, act, and heal themselves. 

-----

Key points covered: 

The evolution of Marvis Actions: How Mist is advancing from anomaly detection to autonomous remediation in network operations 

Architecting for AI-native automation: Key considerations for designing networks that support both driver-assist and self-driving capabilities 

Trust and transparency in automation: How Marvis 2.0 empowers users with visibility and control over AI-native actions 

Operational intelligence at scale: Leveraging continuous data analysis to proactively detect and resolve network issues 

Preparing for agentic AI: How current innovations in Marvis lay the groundwork for scalable, intelligent, and self-healing network architectures 

-----

Where to find Sudheer Matta? 

LinkedIn – https://www.linkedin.com/in/sudheermatta/ 

Where to find Bob Friday?

LinkedIn – https://www.linkedin.com/in/bobfriday/ 

-----

Keywords

Network Automation, agentic AI, marvis actions, marvis minis, self-driving network, AI-powered remediation, AI-driven networks, automatic issue resolution, human intervention, proactive network maintenance, network monitoring, network security sensitivity, network compliance regulations, network insights, Marvis Actions benefits, strategic decision-making in networks, machine learning development, AI networking advancements,  dynamic response  

-----

To stay updated on the latest episodes of The Q&AI Podcast and other exciting content, subscribe to our podcast on the following channels:

Apple Podcasts - https://podcasts.apple.com/us/podcast/the-q-ai-podcast/id1774055892

Spotify - https://open.spotify.com/show/0S1A318OkkstWZROYOn3dU?si=5d2347e0696640c2

YouTube - https://www.youtube.com/playlist?list=PLGvolzhkU_gTogP5IBMfwZ7glLp_Tqp-C

We hope you enjoyed this episode!

Chapters

• 0:00 Welcome and Introduction
• 1:00 Self-Driving Experience Parallels
• 1:40 Current State of Marvis Actions
• 2:24 Marvis Actions Facelift 2.0 Revealed
• 5:30 Future of Self-Driving Networks

Transcript

Bob: Hello and welcome to another episode of Q&AI. Today we're going to be looking at some of the latest innovations around Marvis for Mist, and today we have the honor of being joined by Sudheer Matta, senior VP of Product Management at Mist. Welcome, Sudheer, maybe we can start with a little bit of background for the audience, and how did you end up here at Mist?  

Sudheer: Bob, thank you very much. I'm stunned that this is the first time you invited me to this episode. I've been watching a lot of the episodes. It's been very informative. So, thank you very much. And to the question of how did I end up here. Thank you, Bob. You gave me an opportunity of a lifetime to be at Juniper Mist to run products, and so that's what I do here. So, thank you for being here.  

Bob: Well, welcome. Now maybe we'll start a little bit, because we've been working on this mission to make self-driving networking for the last what 10 years or so? So, I thought maybe we start with. I hear you've had your first self-driving Waymo, Uber experience. What do you think Self-driving, is it all what it's cracked up to be?

Sudheer: Bob, thank you first one more time for leading the way I'd taken a Waymo Uber. I took mine in Arizona, just like you did. It was the most realistic driving experience without a driver I've ever experienced, I mean. So, it was very real, and I think to your point that proves the point that we can make networks, self-driving. 

Bob: So now Marvis Actions is kind of the self-driving component at Mist right now, and this is one of your original babies that you came up with. Maybe give a little description. Where are we with self-driving and Marvis Actions today?  

Sudheer: So, Marvis Actions, basically, Bob, is we are looking at all of the user's data continuously, and we've been historically for the last 10 years we're applying anomaly detection to understand if there is a change and what we were ending up doing was providing actionable insight and where we could take action automatically, we were taking action already. And so, Marvis Actions is basically anomaly detection on all the trends we are seeing and providing AI actionable insight for our customers using that.  

Bob: Okay, now I hear we're going to have a Marvis Action Facelift 2.0 big announcement coming out. So, what is this facelift all about? What am I going to see after my new Marvis Action facelift?  

Sudheer: Bob, I know you're a big champion of this, and so today, as I said, we do self-drive where we can and where we cannot, we driver assist. Basically, we show them and tell them what actions they need to take. Examples are we can't automatically just change a bad cable. That's where we provide driver assist to say, hey, you got to go change the cable. But there are scenarios today where, if we see, imagine a video camera, port sending traffic continuously, we're watching this port all the time and suddenly traffic stops.  

And so, since we are watching the regularity of data, the periodicity of data there and the amount of bandwidth flowing, if suddenly traffic stops, the port didn't go down, the camera is up, power didn't go down, we could auto-remediate that port from being stuck and bring it back. And so that's the self-driving component. So, we have two components driver assist where we can't fix it, they need to take action, but self-driving where we can, and so that's sort of the two categories of Marvis Actions. But you've been imploring us to actually, you know, provide more accountability and gain more trust, and so that's where we are headed.  

Basically, the new facelift for Marvis Actions will say it will show you if an action can be self-driven, and so if it can be self-driven, this is your big thing you've been asking for get trust, gain trust of the user. So, we're going to basically show the action can be self-driven. But the user has a choice. Do you want Marvis to go fix a missing VLAN configuration or a misconfigured Ethernet port? Now we're changing config, Bob, and people get antsy about you know Marvis touching their config right. So, by now saying this action is possible to be self-driven, now the user has a choice, and so that gives them control and as they gain trust, they could say hey, Marvis, you can self-drive.  

Bob: OK. So, it sounds like we're on that verge of really kind of moving from that assisted self-driving where we're giving you some advice to actually self-driving and saying, hey, take your hands off the steering wheel. So, looking at all the pieces and going forward, we got Marvis Minis coming out, large experience model. Where do you think we are? How close are we to that Waymo Uber experience for networking?  

Sudheer: Yeah. So, bob, what we are launching with this announcement is there are several actions that previously were not being self-driven we are putting out there. So changing config, aligning config, fixing stuff so we're moving the ball forward. We're getting ever so close to a self-driving experience. To be honest, there's a lot more we can do, and this is the foundation we are laying for bigger stuff that we are going after with Agentic AI and everything, but all of this is the foundation for that.  

Bob: Sudheer, I want to thank you for joining us. I truly look forward to this self-driving Waymo Uber experience. I want to thank the audience for joining us on this episode and look forward to seeing you on the next episode of Q&AI.