Tradeify Posted "Vibe Coder" on HN. LinkedIn Has Zero of Them.
The May 2026 HN thread made "Vibe Coder" a real title. Title and tool-keyword search miss the pool. Source by shipped artifact instead.
The May 2026 Hacker News "Who is hiring?" thread (item 47975571, posted May 1) did something the recruiting industry has been quietly dreading. Tradeify listed an open role with the literal title "Vibe Coder." Three other posters demanded "top 1% with Lovable, Replit, or Cursor." If you tried to fill those reqs with a LinkedIn title search this morning, you found nothing. The pool exists. Your search primitive is broken.
This is a sourcing problem, not a market problem. The candidates are out there, shipping every day, leaving public artifacts on Replit, Lovable, v0, Bolt, and GitHub. They just don't list themselves under a job title that didn't exist eighteen months ago. If you keep searching by title, or worse, by "X years of Cursor," you will lose every one of these reqs to the founder who DM'd the right person on X last Thursday.
What actually got posted in May
Tradeify's "Vibe Coder" listing is the headline, but it's not an outlier. Scrolling the May 1 thread, the same shape repeats: short job specs, named tool stacks (Cursor, Claude Code, Bolt.new, Lovable, Copilot, Windsurf, v0, Replit Agent), and an explicit ask for speed. "Ship in a week." "0→1 prototypes." "Build internal tools end to end." The aggregator GoodVibeCode tracks the same pattern across hundreds of listings: most vibe coders use two to three of these tools in combination, and the JD reads like a product spec, not a competency matrix.
One representative US listing surfaced via ZipRecruiter is worth quoting verbatim because it captures the entire genre: "tools like Replit, Claude, ChatGPT, Perplexity, Lovable, Cursor, Figma, Canva, WordPress... we're looking for an AI Vibe Coder & Creative Technologist who understands enough code to build and troubleshoot front-end projects, but who also thinks like a marketer, designer, problem-solver, and entrepreneur."
Read that twice. The employer is not asking for a software engineer. They are asking for a builder who happens to use a code editor as one of eight surfaces. A LinkedIn filter on "Software Engineer" + "Cursor" will return the wrong half of the pool, and a filter on "Vibe Coder" will return nobody at all.
Why "2 years of Cursor experience" is a self-own
Cursor's parent company, Anysphere, raised a $900M Series C at a $9.9B valuation in June 2025. The product is barely older than the GA window of most enterprise SaaS contracts. Any JD asking for "2+ years Cursor" is filtering out the actual top-of-pool builders (who picked it up in late 2024 when it stopped being a toy) and filtering in the candidates most willing to round up on a resume.
KORE1's mid-May sourcing guide called this out directly: tool-fluency requirements that age out every six months are not screens, they are noise. Claude Code shipped in production form less than a year ago. Lovable's hiring milestone happened in February 2026. Bolt.new's UX has shifted three times since launch. If your boolean string says "2 years" AND ("Cursor" OR "Claude Code" OR "Lovable"), you are running a filter that mathematically cannot match the people you actually want.
The fix is not to lower the year count. The fix is to stop sourcing on tool tenure and start sourcing on shipped artifacts.
The artifact trail is public, but it's not on LinkedIn
Here is where this pool diverges hardest from a normal engineering search. Vibe coders apply with a URL. Till Freitag's agency, which posted "Germany's first Vibe Coder" job ad as a template, said it plainly in the application instructions: "We want to see what you build. No cover letter, no assessment centre. Just you and your prompt."
That JD is not a stunt. It is a leading indicator of how this market wants to be screened. The application is a deployed app. The screen is a Loom of how it was built. The reference check is a public commit history on GitHub between 2024 and 2026, or a Replit project page, or a v0 community profile, or a Lovable gallery entry.
None of those signals live on LinkedIn. Which is why title search and skill-tag search both miss the pool by something close to ninety-nine percent: the pool was never indexed there in the first place.
The communities that actually contain these people:
- Replit Bounties (paid build history, public)
- Lovable's project gallery (deployed apps, attributed)
- v0.dev community showcases
- Bolt.new showcase pages
- GitHub trending for TypeScript and Next.js, 2025 and 2026 cohorts
- Show HN posts where the author replies in the comments
- The monthly HN "Who wants to be hired?" thread (under-mined; most posters list a portfolio URL)
- X/Twitter build-in-public threads tagged with the relevant tool
Sourcing this pool means crawling those surfaces, joining them on the same person, and ranking by recency and shipped-artifact volume. That is unpleasant manual work in a browser. It is also exactly the problem we built Refolk to handle: you describe the builder in plain English ("ships full-stack React apps using Cursor or Lovable, has shipped at least three deployed URLs in the last 12 months, based in US or remote-friendly EU") and you get a ranked shortlist that joins GitHub, LinkedIn, and the open web behind a single query.
The application is a deployed app. The screen is a Loom. The reference check is a commit graph.
The Palantir FDE diaspora is the closest known proxy
If you have to start from a known title bucket (because your ATS demands one), the right bucket is not "AI Engineer." It is Forward Deployed Engineer and Growth Engineer. Refolk's index shows roughly 1,464 US profiles across those two adjacent titles, concentrated in NYC and SF, with top current employers reading like a who's-who of the AI-native operator stack: Palantir, Salesforce, Modal, Cursor, ElevenLabs, Scale AI, and Beacons.
This is the talent loop to mine. Palantir's FDE archetype (deploy on-site, build the integration in a week, own the customer outcome) maps almost perfectly onto what Tradeify and the rest of the May HN thread are buying. The diaspora has already scattered, and the people who left Palantir for Modal or ElevenLabs in 2023-25 are exactly the ones now picking up Cursor and Claude Code as force multipliers on the same workflow they were already running.
If you have to write one boolean, write this one: ("Forward Deployed Engineer" OR "Growth Engineer" OR "Solutions Engineer") AND (Palantir OR Modal OR ElevenLabs OR Scale OR Cursor OR Beacons) AND active GitHub commits 2025-26. You will hit something closer to fifteen percent of the real pool, which is fifteen times better than a title search on "Vibe Coder."
Non-engineers are competitive, and your filter is hiding them
Lovable's first professional vibe coder hire, profiled in Lenny's Newsletter in February 2026, was explicitly framed as a non-technical professional shipping enterprise-grade products without writing code. Lazar Jovanovic was not a CS major. He was not a senior SWE who learned to prompt. He was a builder who happened to use the tool the company makes.
This is the part most recruiters refuse to internalize. If you require a CS degree, you cut the strongest portfolios. If you require "Software Engineer" as a prior title, same outcome. The competitive applicants in this pool are routinely ex-designers, ex-marketers, ex-PMs, and ex-founders whose GitHub is thin but whose deployed-URL count is double digits.
Cortance, a staffing shop pitching this exact pool, framed the real screen well: "the most common bottleneck when hiring vibe coders is not finding someone who uses AI tools. It is finding one who knows exactly when not to trust the output." That is a judgment screen, not a credential screen. You cannot run it through a resume parser. You run it by looking at what the person actually shipped, what broke, and how they handled it in the commit log or the Show HN comments.
The compensation has caught up. The titles have not. The sourcing tools mostly have not. That gap is the opportunity.
A concrete playbook for the next two weeks
If you have a Tradeify-shaped req open right now, here is the order of operations that actually works.
- Throw out the title. Rewrite the JD around outcomes: "ship a deployed prototype in week one, own three to five internal tools by month three." This is what the May HN thread is buying and what the candidates respond to.
- Define the artifact bar, not the tool bar. "Three deployed URLs in the last twelve months" is a real screen. "Two years of Cursor" is not.
- Pull the FDE/Growth Engineer adjacent pool first. That is the fastest fifteen percent. Palantir, Modal, ElevenLabs, Scale, Cursor itself.
- Crawl the artifact surfaces. Replit Bounties leaderboards, Lovable gallery, v0 community, GitHub trending, Show HN authors with active 2026 commit graphs. This is where the other eighty-five percent lives, and where tools like Refolk earn their keep by joining a Replit profile, a GitHub handle, and an X account back to one human you can actually email.
- Screen on judgment, not credentials. Ask for a Loom walkthrough of one project the candidate is proud of and one they would rewrite. The answer to "what would you rewrite" tells you everything Cortance's framing is asking about.
- Move in days, not weeks. The May HN thread will be answered by the end of June. The same candidates will be picking between three offers by then. If your loop is four weeks of panels, you are not in this market.
The headline shift is simple. "Vibe Coder" is a title that means nothing on LinkedIn and everything on a deployed URL. Source the URLs. The humans come attached.
FAQ
Is "Vibe Coder" actually a durable job title, or will it disappear?
The title will probably get renamed within a year (most likely back into "Growth Engineer" or "Product Engineer"), but the role is durable. What's underneath is a builder who closes the gap from idea to deployed prototype in a week using AI tools as primary surface. That workflow is not going away because Lovable, Cursor, and Claude Code are not going away. Source the workflow, not the noun.
How do I screen for "knows when not to trust the output" without a take-home?
Ask for a fifteen-minute Loom of a project the candidate built and a five-minute Loom of one they would rewrite. The rewrite Loom is the screen. Candidates who say "I'd add tests here, I caught the model hallucinating a Stripe API that doesn't exist, I refactored auth myself because the generated version had this specific flaw" pass. Candidates who say "it works, I'd ship more features" fail. Total candidate time: twenty minutes. Total recruiter time: thirty minutes. No take-home required.
My ATS requires a job title for sourcing. What do I put?
Use "Growth Engineer" or "Forward Deployed Engineer" as the system-of-record title and write the real spec into the JD body. That gets you the ~1,464 adjacent-bucket profiles in the US to start with, and lets you layer artifact-based screening on top. Avoid creating a new "Vibe Coder" req category in your ATS until the title stabilizes; you'll just create reporting confusion six months from now.
Where do I actually find these people if I only have an hour?
Start with the May 1 HN "Who is hiring?" thread (id 47975571) and the same month's "Who wants to be hired?" thread. Then Replit Bounties top earners 2025-26. Then GitHub trending for TypeScript and Next.js, filtered to accounts with active 2026 commits and a linked personal site. One focused hour gets you twenty to thirty real names. Joining those identities to email addresses is the slow part, which is what a sourcing tool that crawls GitHub, LinkedIn, and the open web in one query is for.