Do you need AI on your CV now?

“Learn AI” is the loudest advice in tech. We counted which job ads actually ask for it.

By the Glassbox team · the week of 19 October 2026

This story keeps to one week, 19 October 2026. The AI page starts a new count every Monday.

A wooden door standing open in an old stone wall.

This week, Holzland Becker, a German timber and building supplies company, advertised a student software job.

It asks for PHP, JavaScript, Python, and LLMs.

If a timber merchant wants AI, does everyone? We counted.

One door for each job family we count. From here on, we draw them.

Door photos: Pexels, credited at the end · swipe along the row

Out of every 100 software engineering jobs advertised this week, how many name LLMs?

Drag up the building to light its windows, or tap one. Or keep scrolling, and we’ll show our count anyway.

Your guess none yet

Each window is 1 of every 100 software engineering jobs advertised this week. The windows you light get a pen ring. Our count, 16 of the 100, lights lime. LLMs means the ad names large language models, or a close term on our list, as required or preferred.

Source: Glassbox count, week of 19 October 2026 · LLMs: the ad names large language models or a close term on our list · how we count

It depends which door you walk through.

62% of AI jobs name LLMs (494 of 800). In machine learning, 45% (207 of 455).

In software engineering, 16% (523 of 3,370). In DevOps, 8% (56 of 743). In embedded jobs, 1% (5 of 366).

Whichever door is yours, curious which AI skills are already on your CV?

Your CV is probably on your laptop. Comparing it there takes a minute.

Graphic: a street of 19 drawn buildings, one per job family, sorted from the loudest ask for LLMs to the quietest. Every window holds a drawn person for 5 jobs, rounded; the windows of jobs that name LLMs are lit lime. Each sign carries its share and its count: AI 62% (494 of 800), machine learning 45% (207 of 455), data science 28% (74 of 269), solutions 23% (298 of 1,319), data engineering 22% (71 of 322), full stack 21% (76 of 364), security 17% (97 of 563), backend 16% (67 of 413), software engineering 16% (523 of 3,370), product 13% (191 of 1,454), IT support 11% (60 of 553), data analytics 10% (20 of 199), design 10% (49 of 503), frontend 9% (17 of 188), QA 9% (32 of 357), DevOps 8% (56 of 743), mobile 6% (10 of 165), embedded 1% (5 of 366), game 0 of 15.

Source: Glassbox count, week of 19 October 2026 · LLMs: the ad names large language models or a close term on our list · how we count

Here’s the odd part. 47% of AI jobs don’t name a single programming language (379 of 800).

Elastic’s Marketing AI Engineer asks for LLMs and technical writing. JavaScript and Python are only nice to have.

Graphic: the AI door opens and its 160 drawn people step onto the pavement in four groups, one laptop per programming language their ad names: 47% of AI jobs name none (379 of 800), 25% one (196), 17% two (134), 11% three or more (91). The Elastic listing is pinned on the door.

Posted 14 Oct 2026

Marketing AI Engineer

Elastic

Bangalore, India

On site4 years

Requirements

Elasticsearch
observability
webhooks
LLMs
workflow automation tools
technical writing
Google Workspace

Nice to have

JavaScript
Python
accessibility

Skills as we read them from the ad

Source: Glassbox count, week of 19 October 2026 · LLMs: the ad names large language models or a close term on our list · how we count

AI is leaking into jobs that aren’t about AI. Airbnb’s “Operations Engineer, BizTech” lists 26 required skills. Two of them are AI: LLMs and AI agents.

Graphic: a notice board on a lamp post at the end of the street, between the doors, with the Airbnb listing pinned on it: Operations Engineer, BizTech, 26 required skills, LLMs and AI agents lit lime.

Posted 18 Oct 2026

Operations Engineer, BizTech

Airbnb

Anywhere in the World

Remote3 years

Requirements26 required skills

incident response
LLMs
AI agents
data quality
Tableau
observability
root cause analysis
Python
Go
JavaScript
TypeScript
Bash
Elasticsearch
SQL
Airflow
message queues
Terraform
CI/CD
Jenkins
GitHub Actions
AWS
GCP
Jira
Salesforce
Workday
identity and access management

Nice to have

None listed

Skills as we read them from the ad

Source: Glassbox count, week of 19 October 2026 · LLMs: the ad names large language models or a close term on our list · how we count

And it’s asked more of senior engineers than of beginners: LLMs in 11% of entry level software jobs (72 of 627), and in 17% of senior ones (485 of 2,796).

Graphic: two doors, entry level and senior. In front of the entry level door, 125 drawn people for 627 jobs, 14 lime for the 72 that name LLMs. In front of the senior door, 559 drawn people for 2,796 jobs, 97 lime for the 485 that name LLMs.

Source: Glassbox count, week of 19 October 2026 · LLMs: the ad names large language models or a close term on our list · how we count

Where it’s quiet. 1% of embedded jobs name LLMs (5 of 366).

Qualinx’s Senior Embedded Software Engineer asks for C, Python and embedded systems. Nothing AI.

Graphic: the embedded door opens; its 73 drawn people stand on the pavement, one of them lime with a pen ring: 1% of embedded jobs name LLMs, 5 of 366. The Qualinx listing is pinned on the door.

Posted 18 Oct 2026

Senior Embedded Software Engineer

Qualinx

Delft, Zuid-Holland, Netherlands

On site10 years

Requirements

embedded systems
C
Python

Nice to have

None listed

Skills as we read them from the ad

Source: Glassbox count, week of 19 October 2026 · LLMs: the ad names large language models or a close term on our list · how we count

So, do you need AI on your CV? Some doors ask loudly, others barely mention it.

Graphic: the whole street again, every building with its share and count; a woman stands at the AI door, a man at the embedded door.

Source: Glassbox count, week of 19 October 2026 · LLMs: the ad names large language models or a close term on our list · how we count

Posted 18 Oct 2026

Operations Engineer, BizTech

Airbnb

Anywhere in the World

Remote3 years

Requirements26 required skills

incident response
LLMs
AI agents
data quality
Tableau
observability
root cause analysis
Python
Go
JavaScript
TypeScript
Bash
Elasticsearch
SQL
Airflow
message queues
Terraform
CI/CD
Jenkins
GitHub Actions
AWS
GCP
Jira
Salesforce
Workday
identity and access management

Nice to have

None listed

Skills as we read them from the ad

Source: Glassbox count, week of 19 October 2026 · LLMs: the ad names large language models or a close term on our list · how we count

Which door are you standing at?

AI is a skill some doors ask for loudly and others barely mention.

One week of job ads, counted door by door: that is where LLMs are named, and where they barely appear.

Curious which AI skills, LLMs among them, are already on your CV?

Your CV is probably on your laptop. Comparing it there takes a minute.

One week, 19 October 2026. The AI page counts again on Monday.

What we can’t see

A named skill isn’t a hiring rule. We count what ads name, required or preferred.
Our boards lean towards AI companies, which pushes these shares up.
“LLMs” means the ad names large language models or a close term on our list, not that the job builds them.
One week’s count. Not a trend.

How we counted

Each building is one job family. The job type comes from the title. Senior means a senior, lead or principal level in the same job type.

LLMs is one skill on our list: the ad names large language models, or a close term such as generative AI or prompt engineering, whether as required, preferred or in passing. Every count here is of the ads long enough to read, and one drawn person is 5 of them, rounded.

The two doors by level count the software jobs on the entry level page. Entry level means one of three things. The job's level reads as intern, entry level or junior. Or its title says intern, graduate, new grad, entry level, early career or junior. Or it requires 2 years of experience or less and its level isn't senior. Software means five job types: software engineering, backend, frontend, full stack and mobile.

The jobs come from public job boards, read every six hours: employers' own boards on Greenhouse, Ashby, Lever and Workable, and job sites such as Remote OK, Remotive, Arbeitnow, Jobicy, Himalayas and We Work Remotely. Each job counts once, however many boards or cities list it.

What we can't see: jobs on boards we don't read, and jobs never posted in public; skills a job names in words that aren't on our list; whether a job is filled, how many people apply, or who gets hired.

How Glassbox counts.

Door photos: Pexels, by Nati (the wooden door), Guru Saran, Juan Martin Lopez, Florian Süß, Ellie Burgin, Burcu Elmas, Viridiana Rivera, Shiva Kumar Reddy, Mieke Boogert, Domenico Bandiera, Pixabay, Damir K, The Daphne Lens, Olga K, G N, Polat Eyyüp Albayrak, Jan van der Wolf, Rendy Tomaluweng and Eren Ataselim. Photo heads: Pexels, by Andrea Piacquadio and Ventara Films. Drawings: Glassbox. Sign lettering: Tilt Neon (OFL).