The Evolving Role of Streaming Engineers in the Age of AI and Automation
Streaming engineers’ traditional functions are increasingly usurped or at least altered by the offloading of observability and monitoring and much decision-making to automated solutions. When it comes to streaming infrastructure management, how much have AI and automation driven humans from the loop? Where are human eyes on glass—however comparatively inefficient—still critical to sound streaming operations? Do automated solutions actually enable engineers to accomplish more than they could before? Servers.com Senior Global Sales Executive Lukas Navickas, AI & Streaming Technology Strategist and SVTA AI Committee Co-Chair Bhavesh Upadhyaya, Qualabs CEO Juan Pablo Saibene, and LiveU Solution Specialist Ben Ratner discuss in this clip from their panel at Streaming Media Connect 2026.
Keeping the Human Touch
Upadhyaya directs his first question to Navickas, raising the issue of a possible changing relationship between streaming operators and engineers and the streaming infrastructure they manage due to an influx of new AI-powered and automation-oriented tech: "Do you have tools now that are allowing one engineer or operator to do more with managing the infrastructure than they could have five years ago?" he asks.
"So I think this question is twofold, right?" Navickas replies. "If we go back to that physical layer of actually running and operating a data center, from that perspective, as we scale as a platform or as we scale based on acquisition, then we tend to see increased requirement for engineers to actually go at the data centers, rack and stack, prepare and look after what we're trying to achieve. And of course at the operational level," he continues, "when it comes to overall automation, networking, and connectivity, we also see a lot of increase in our engineers. So it kind of does scale for us even though, as mentioned there, in-built tools of AI, it do definitely enable our software-layer engineers to work on automation tools, solve challenges easier."
But he goes on to describe Servers.com's "own ethos" as built around a preference "to keep that human touch alongside a customer journey rather than push them through a ticketing system, or, 'Hey, we've got an issue with a live stream,' or, 'Hey, we're spotting an issue with one of our end users.' So we really tend to appreciate that human touch where our customers do really hands-on. So we have a mixed bag of emotions around this idea of completely automating everything from an AI perspective."
Observability First
Saibene acknowledges the value of a human presence in the observability loop for live streaming but maintains that AI tools play an essential role in augmenting what humans can do in the volume of processing necessary to monitor everything that's happening throughout the pipeline. "First we need observability, and that's the data," he says. "We need to understand what's going on across the stack, what's happening in the pipelines, what's going on with the distribution, CDN, the ISPs, and also what's happening at the client. But then processing those millions of logs you're getting every few seconds--it's impossible for a human to go through that."
Noting that Qualabs has "had multiview dashboards for a long time," he says that more recently, "we've been working a lot on agents and AI-enabled tools that actually augment the operator's ability to scale. But we have one rule: Our runtime decision path is deterministic, but we can add probabilistics into the root-cause loop. So we do have tools that learn faster and augment the humans' capabilities to operate. But under pressure when you're doing it, you've got to have it rehearsed, you've got to know your playbooks, your runbooks, and you need to understand what you're doing because there are 100 moving pieces and everything could go wrong."
"I think observability is across the entire stack," Upadhyaya chimes in, "because once you have a human operator in this pipeline doing more, they can't keep track of everything. How do you build the tools that make everything observable and make it actionable so that you're not paying attention to every single screen?"
Two Kinds of Monitoring
Upadhyaya then turns to LiveU's Ratner for more of a production-side perspective and says, "Ben, you've got a multiview, you're putting four cameras out there, right? Let's talk about what good observability looks like on your platform. How do you give the right tooling to the human beings to allow them to do more confident that they're not screwing things up?"
"There's two kinds of monitoring that you need to do when it comes to the production side of things," Ratner replies. "There's quantitative monitoring, which is, 'Is the signal coming in? Do I have the right bandwidth? Is it going out the right way?' The physical infrastructure piece. And then there's the quality piece of it. Quality is in the eye of the beholder. 'Does it look good? Is it doing what it's supposed to do? Is it getting the purpose of this production done?' The former, the quantitative part, is a lot easier to do. You can aggregate numbers into views very easily. We actually have a tool called Actus that helps with some of that, as well as in LiveU Studio, you can see a lot of that. But when it comes to the quality of the content, the actual output, that's something it's much more helpful to have human eyes on."
Conceding that today's humans-in-the-loop status quo won't necessarily last forever, he says, "Now, this is not to say that in the future our good friend AI won't be able to do a better job at this, but right now at least you need to have eyes on that. So that's where scale hits its ceiling, for now at least."
Related Articles
Streaming Media Connect returned this week with six dynamic keynotes from NASA+, Google, TiVo, BallySports, Wowza, and TwelveLabs, and peak-form panels on refining your live streaming playbook, tuning latency, matching ads to content, scaling streaming architecture, curbing churn, leveraging AI in personalized content discovery, innovating streaming UX, optimizing QoS and QoE, and a timely dissection of Disney's 4K EU blackout.
14 Aug 2026
The World Cup is live. The infrastructure is either holding or it isn't. The difference was decided long before the opening whistle.
24 Jun 2026
Cloud and remote production workflows have changed the game for live streaming, enabling significant CapEx and OpEx reductions, making it possible to achieve the same results with smaller and more efficiently deployed and largely off-site crews and reducing hardware footprints. But do all of the gains in operational efficiency that remote and distributed production provide come at the expense of real-time observability? It definitely brought new challenges on that end, according to Telestream director of product management Ken Haren, who explores these challenges and changes in best practices for software-defined observability and live-stream diagnostics that have come with the shift to remote production in this discussion with Zixi SVP of business development Emeka Okoli and Eyevinn Technology media solution specialist and VP of sales and business development Magnus Svensson at Streaming Media Connect 2026.
01 Jun 2026
As with all streaming workflows, AI has steadily crept into the live streaming technology stack. In some cases, the impact is incremental, in others, profound. From production to monetization, here's a quick overview of where AI has become relevant for live event producers and engineers, and some areas where, surprisingly, it hasn't.
26 Mar 2026