Refolk
10 min readMarkdown

LLM or Claude on Your Resume? What PM and Designer Postings Ask For

LLM tops PM postings and Claude ranks 2nd for designers. How to list AI skills on a resume as outcome bullets that hold up to an interviewer's follow-up.

If your resume says "LLMs" and the posting says "Claude," a recruiter searching for the posting's terms may never see you. If you add "Claude" to your skills list with nothing behind it, one interview question about it is enough to knock the claim down.

Lenny Rachitsky's State of the Tech Job Market report, published Nov 10, 2026, shows how much that wording now matters. "LLM" is the most-named skill in product manager postings, "Claude" ranks third for PMs and second for designers, and every role now asks for AI skills. Below: where those terms belong on a resume, and why the exact tool name matters more than the general one.

What did Rachitsky's late-2026 report find about AI skills?

AI tool names have pushed past the old technical staples near the top of PM and designer job descriptions, and each role names a different one. PM postings lead with the general term "LLM." Designer postings name "Claude" itself.

This is Rachitsky's fifth market report. It draws on two datasets:

  • TrueUp, which tracks global job openings at more than 9,000 tech companies.
  • Live Data Technologies, which tracks hires, departures, and promotions for 124 million US professionals.

Two years ago, the most-mentioned skills in PM job descriptions were SQL, Jira, AWS, and Python. Now "Claude" is the third most-mentioned term, and Cursor, OpenAI, and MCP are all in the top eight. For designers, Claude passed CSS this spring to become the second most-mentioned skill. Rachitsky points out that the postings say Claude specifically, not "LLM" or ChatGPT.

A third-party reading of the paywalled charts (from cellcog.ai, a publisher that runs on Claude, so treat it as secondary) gives these shares:

  • PM postings: LLM 11%, SQL 9%, Claude 5%, AWS 4%, Cursor 3%.
  • Designer postings: Claude 13%, Cursor 7%, CSS 7%, LLM 6%.
  • Engineering postings: AWS 19%, Kubernetes 13%, LLM 12%, Java 11%, Claude 6%. LLM has just passed Java.

Where is the gap between postings and candidates?

The gap is the specific tool name. Postings ask for "Claude" far more often, relative to "LLM," than PMs mention it when they describe themselves, while the general term "LLM" is already everywhere.

In Refolk's index of professional profiles, 834 US people with a "Product Manager" title have "LLM" in their headline. Only 13 have "Claude." So among PMs who put either term in their headline, about 1.5% chose Claude.

Role / termShare of postings naming itUS profiles with term in headline (Refolk's index)
PM, "LLM"11% (#1)834
PM, "Claude"5% (#3)13
Designer, "Claude"13% (#2)not queried
Designer, "AI" (generic)n/a765
Engineer, "LLM"12% (above Java at 11%)not queried

Posting shares are cellcog.ai's reading of Rachitsky's charts. Profile counts are from Refolk's index, matching on headline keyword.

Put the ratios next to each other. In PM postings, LLM outnumbers Claude about 2.2 to 1. In PM headlines, it outnumbers Claude about 64 to 1.

29x
How much more often PM postings name Claude, relative to LLM, than PMs do in their own headlines

Postings run LLM to Claude at about 2.2 to 1. PM headlines in Refolk's index run 64 to 1.

Headlines aren't resumes, so treat this as a proxy. The direction is still clear. PMs adopted the category word and skipped the brand name, and the brand name is where postings have moved.

Why does the exact tool name matter so much?

The exact name matters because recruiters search applicant tracking systems for the posting's own terms. If the posting says "Claude" and your resume only says "LLMs," that search won't connect them.

So use the posting's exact phrase, in the posting's form. If the job says "generative AI," write "generative AI," not "gen AI" or "LLMs." The same goes for Claude, Cursor, and MCP.

There's also a positioning reason. With 834 PMs already writing "LLM" in their headline, that word doesn't set you apart anymore. It's the minimum. Naming the actual tool, in a bullet that shows what you did with it, is what moves you from "familiar with AI" to "has shipped with this."

Why is naming a tool a liability as well as a keyword?

Every tool you name invites a follow-up question. A specific name gets tested harder than a vague one, so it only helps when a real result backs it up.

Recruiters increasingly ask directly about any AI tool on a resume: what you built, what went wrong, what you'd change. "AI-savvy" rarely draws a pointed question. "Claude" gets "show me." If all you have is "I've used it in chat," the name you added to pass the screen now works against you in the interview.

The vague version is still the norm. Roughly 85% of resumes that mention AI do it in terms like "AI enthusiast," according to a JobSprout figure quoted secondhand. Robert Half's 2026 survey shows the cost from the other side:

65%
HR leaders who struggle to confirm skills on AI-tailored resumes

In the same 2026 Robert Half survey, 67% said these resumes delay hiring.

That's the overfitting trap. A resume that copies every term in a posting gets a strong keyword match and then turns out to have weak proof in the interview. The fix isn't fewer keywords. It's putting each keyword next to a result.

A bare tool name in your skills section is a question you haven't answered yet.

How should you write an AI skill into a bullet?

Name the tool, then the task, then a measurable outcome, all in one experience bullet. Then repeat the tool name once in the skills section for keyword matching.

The formula:

  1. Tool: the exact name the posting uses (Claude, Cursor, MCP, "LLM").
  2. Task: what you actually did with it, phrased as work, not as familiarity.
  3. Outcome: a number you can defend, such as time saved, error rate, adoption, or revenue.
  4. Skills line: the same tool name once more, so the keyword search finds it.

ResumePulse publishes a good model: "Used Claude and internal LLM tooling to automate weekly reporting, cutting a 6-hour manual task to 40 minutes." It names the tool, describes the task, and gives a before and after. An interviewer's follow-up ("walk me through the automation") has an obvious answer.

Here's the same experience written badly and well:

VersionWhat it saysWhat an interviewer asks
Skills dump"AI tools, LLMs, prompt engineering""Which tools? For what?"
Vague bullet"Leveraged AI to improve team efficiency""How, and by how much?"
Outcome bullet"Used Claude to draft and triage [N] support tickets weekly, cutting first response from [X] to [Y]""What broke along the way?" (a question you want)

Use your own real numbers in the brackets. If you don't have a number, describe the scope: how many people used it, which workflow it replaced, what shipped because of it.

Doing this by hand for every application adds up, because the right tool name changes from posting to posting. That's the work Refolk takes off your plate: paste the posting, and Refolk rewrites your resume from your own history using the posting's terms, keeping the bullets tied to things you actually did.

What AI skills belong on a product manager resume?

A PM resume should use "LLM" as the baseline term, since it's in 11% of PM postings, and add Claude, Cursor, or MCP only where you've shipped or run something with them. LLM experience on a resume is what's expected now. The named tool is what sets you apart.

The market makes this worth doing. PM openings are up 75% from their March 2023 low, and there are now 3.8 open PM roles for every open design role, up from 1.9 in early 2023. More openings means more postings to tailor to, and each one weights the vocabulary a little differently.

Good PM bullets connect the tool to a product decision:

  • Shipped a feature built on an LLM: name the model or vendor if the posting does, and give the adoption or retention number.
  • Prototyped with Cursor or Claude: say what the prototype proved and what happened next (greenlit, killed, pivoted).
  • Ran evals or set guardrails: describe the failure rate you started with and where you got it to.
  • Wired up MCP integrations: name the systems you connected and the workflow they replaced.

Skip the bullet that says you "used ChatGPT to write PRDs faster." It names a tool and no result, and the interviewer will ask what was wrong with the PRDs.

How should designers list AI tools on a resume?

Designers should name Claude specifically when the posting does. In designer postings, Claude (13%) shows up roughly twice as often as CSS (7%) or the generic "LLM" (6%), and AI tools on a designer resume work best as process bullets: what you prototyped, tested, or shipped faster because of the tool.

Designers face the tightest version of this problem. Product designer openings fell 28% over the past year to a record low. With fewer openings, each application has to match its posting more closely, and designer postings reward the brand name over the category.

Designers are already moving their headlines. In Refolk's index, 765 US product designers have "AI" in their headline. The employers named most often in that group include Meta (3) and Microsoft AI (2), and new hybrid titles show up, such as "AI Product Designer" and "AI Design Generalist." Note the word they chose: "AI," the generic term again. That leaves room for the designer who writes "Claude" in a bullet with a result attached.

Designer bullets that hold up:

  • Prototyping speed: "Built interactive prototypes with Claude and Cursor, cutting concept-to-test time from [X] to [Y]."
  • Research synthesis: the number of interviews or tickets synthesized, and the decision it fed into.
  • Designing AI features: if you designed the interface for an LLM product, say so in the description of the work, not only in skills.

What about engineers?

Engineers should treat "LLM" as a core skill next to their language and cloud stack, because it just passed Java in engineering postings (12% vs 11%). Claude appears in 6% of them, so name it when the posting does.

The broader numbers point the same way. AI engineer openings rose 46% over the past year, and forward deployed engineer roles grew 4.6x. For engineers, the outcome is usually technical: latency, cost per call, eval pass rate, or the number of internal teams using the service. "Integrated Claude via API" is a skills-list line. "Moved [workflow] onto Claude via API, cutting cost per request by [X]%" is a bullet.

Does the skills section still matter?

The skills section still matters, but only as a secondary keyword match. Put each tool name there once so the search registers it, and make sure every name there also appears in at least one experience bullet.

A quick audit:

  • In skills and in a bullet: keep it.
  • In skills but no bullet: write the bullet or cut the term.
  • In the posting but nowhere on your resume: add it only if you can answer "what did you build with it?"
  • On your resume but not in this posting: fine to keep, but don't lead with it.

The last point is where tailoring comes in. A PM posting heavy on Cursor and MCP calls for different bullets at the top than one that only says "LLM." Refolk handles that per posting, and it also scores how well you actually fit. If a posting asks for MCP and your history has no MCP, you'll see the gap before an interviewer finds it.

FAQ

Should I write "Claude" or "LLM" on my resume?

Use whichever term the posting uses, in its exact form. For PM roles, "LLM" is the most common term (11% of postings) and Claude appears in about 5%. For designers, Claude is in 13% of postings and "LLM" in only 6%. If you've done real work with Claude, name it in a bullet and put "LLM" in skills, so you match either search.

What if I've only used Claude or ChatGPT for everyday writing?

Then don't make it a headline skill. Look for a workflow where the tool changed a result: a report you automated, research you synthesized, a prototype you shipped. If you can't find one, build a small project before you claim the tool, because the interviewer will ask what you built with it.

How often should I change which AI tools I name?

Every application. The posting decides the vocabulary: one PM role leans on Cursor and MCP, another just says "LLM." Reorder your bullets so the tools that posting names come first, and use its exact phrasing so a recruiter's search finds you.

The data behind this

Put this to work

Paste your career in once. Every application after that is written for you.

Drop a resume or a LinkedIn URL. I rank the live openings against it, rewrite the resume and write a cover letter for the best of them, and fill in the employer's form when you press the button. You read, you decide what goes out.

  1. 01Drop your resume

    A PDF or a LinkedIn URL. About a minute, once.

  2. 02I rank the openings

    Every weekday morning, the live catalog scored against your history. Up to 20 worth your time, not two hundred links.

  3. 03Each one is written up

    Resume rewritten for the posting, a cover letter, a fit score. Press send, or let me fill in the form.

  • New matches ranked and written before you are up.
  • Every bullet stays inside what your history supports. Nothing invented.
  • Queued, submitted, interviewing, offer: one screen, not a spreadsheet.

Free, no card. Up to 25 applications sent a week, and nothing is sent until you say so.

Keep reading