As of November 1, 2026, Level lists 582 live jobs that match site reliability. Of these, 485 are Little AI.
Only 68 sit at Works on AI or Builds AI.
Why SRE and AI are in the news
A February 2026 essay from observability vendor SigNoz reached the Hacker News front page this week with 650 points. Its title says AI is not replacing SREs, it is deskilling them.
The author leans on "Ironies of Automation", a paper by Lisanne Bainbridge. The idea is that the more a system automates, the rarer and harder the cases left to the human become.
The essay lists what is already automated in SRE work, including alert grouping, runbook execution for known issues, and log pattern detection. It credits PagerDuty with a three-tier model. Known issues are fully automated, and partly understood ones get AI recommendations with human validation. Novel, cascading failures stay human-led, with AI providing supporting context.
On the vendor side, Rootly sells "AI SRE" software and publishes a guide to it. The tool investigates a production incident alongside the on-call engineer. Its own guide calls it an aid, not a substitute for SRE engineering.
Search interest follows. In Google Trends for the US between October 25 and November 1, "sre ai" averaged 83 on the relative index. "sre jobs" averaged 41. The index is relative and shows interest, not intent.
How many site reliability jobs are open, and how many involve AI?
As of November 1, 2026, Level lists 582 live jobs that match site reliability. They come from 180 employers.
At Little AI (level 1), there are 485 jobs. That is about 83% of the total.
At Uses AI (level 2), there are 29 jobs.
At Works on AI (level 3), there are 40 jobs.
At Builds AI (level 4), there are 28 jobs.
Together, Levels 3 and 4 hold 68 jobs. That is about 12% of the total. Our level definitions explain how each job is placed.
What do the Level 3 and 4 jobs look like?
SpaceX has 12 of the 68 roles, out of its 47 site reliability jobs. Mistral AI follows with 6.
Senior is the most common seniority, with 39 roles. Lead has 11 roles.
On tools, Kubernetes appears in 47 of the 68 postings. Python appears in 19, and Terraform in 18.
Of the 68, 25 are remote. That is about 37%, against about 23% across all 582 jobs.
Of the 68, 36 list a salary in US dollars. Their median advertised minimum is $165,000, a gap of $13,500 with a small sample.
For comparison, the Little AI jobs show a median advertised minimum of $151,500. That figure covers 164 postings with a US dollar salary.
Which companies hire the most site reliability engineers?
SpaceX has 47 matching roles, the most of any employer. SpaceX is followed by LSEG with 32.
Okta lists 22. NVIDIA and MongoDB list 18 each.
You can browse every employer on the companies page.
What do these jobs pay, and can you work remotely?
Across all 582 jobs, 204 list a salary in US dollars. The median advertised minimum is $152,500.
Remote roles number 131 out of 582. That is about 23%. For more, start from our remote AI jobs page.
A median of advertised minimums is not a pay floor. It describes what employers chose to publish.
What does the deskilling debate mean for a job search?
The SigNoz essay suggests practice is the issue. If tools handle routine incidents, engineers see fewer of them. Rare incidents then land on people with less hands-on time.
The practical read is to check what each posting asks for. Check the stack it names, such as Kubernetes, Python or Terraform. Then check whether it expects you to work with AI tools or on them.
