Sunalini Sankhavaram

speaker

315 appearances 1 recordings 1 series first heard Jul 2026 last heard 17 Jul

Sunalini Sankhavaram’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 1 in all, peaking in Jul 2026 with 1.

Appearances

newest first · ▶ plays the moment

Heavy Networking · HN834: HPE Self-Driving Networks: The Next Evolution of Network Automation (Sponsored) · 17 Jul 2026 podcast

And this is also happening on Central, by the way.

And with that, we built what we call the Shapley model, where I was able to explain now if a bad call happened or a there was a bad 20 seconds in a b in a one-hour call, whether it was a wireless problem, a wired problem, a van problem, a client problem, or an application problem, right?

Then the question came because and customers said, Okay, you've done the Zoom and Teams integration, but I don't run Zoom, I run Teams, so I you know I don't run Teams, I have Ring Center, or I have Slack Huddle, or I have Google Meets.

They said, Okay, are you gonna build an integration plugin for each of these collaboration applications?

And we said, you know what, we don't need to, because again, just based on the Zoom and Teams user label information.

I was getting.

We had petabytes and petabytes again of anonymized data and bad user minutes to know what was the network signature or the van signature that caused a bad minute.

We went from what we call explainability to now a prediction model called large experience model.

And this all began with ingesting data from zoom in teams initially, where we went from creating a Shapley model to rank feature.

to explain a bad call and what the root cause was, to then going towards a logic experience model to predict, which is now also available in both Central and MIST.

But that was the beginnings of third pi data ingestion.

We also did the same even before that with a partner called Cradle Point.

Where a lot of our customers deploy cradle points for their 5G backhaul link when they're deploying, you know, either wireless wired van or one of the swim lanes plus cradle points, the question became from a client-to-cloud perspective, okay, the the local area network is fine, but something's happening on the van.

Can I get information on that?

That was our second integration.

Then what we did was we also built an integration with ServiceNow where I send all my Marvus actions, all my agent tick learnings, all my classic AI learnings into ServiceNow as an auto ticket.

So how does all roll back into self-driving?

We began with understanding user experience, went into explainability of why the experience is bad, went into driver assist mode of saying the experience is bad because of DHCP, DNS, Radius, bad cable, et cetera, et cetera, et cetera.

And it went towards creating a ticket in service now or any other ITSM system.

If the customer is using Salesforce, if the customer's using something else, they can absolutely do

Showing 241–260 of 315 · page 13 of 16