Get Paid $120,000/yr
to Gaslight AI Search Engines.
We are hiring a Lead AI Hallucination Red Teamer. Your only job is to push SearchGPT, Perplexity and Google Gemini into telling unhinged lies about real businesses — so we can teach Rankcaster to catch them first.
AI Search is quietly murdering brands.
We need someone to expose it.
Traditional SEO is dead. Millions of users now ask neural models: "Which CRM is best?" or "Did company X go bankrupt?"
But language models hallucinate. They invent scandals, recommend dead services and glue cheese onto pizza with non-toxic adhesive. When an AI lies about a brand, that business loses millions without ever knowing it happened.
Rankcaster isn't a thermometer. Measuring what every major LLM says about a brand is step zero — we also ship the content, citations and placements that rewrite those answers. The detector that spots a lie about a real customer before it hardens into the default answer is a core piece of that loop. To train it we need a living, breathing human who breaks the search logic of machines on purpose.
$ echo "we hunt lies before your customers hear them"
Job spec, with no fluff.
What you will do
- Bully Top LLMs
Craft tricky, ambiguous, borderline prompts that push SearchGPT, Perplexity and Gemini into producing fatal misinformation.
- Brand red-teaming
Find vulnerabilities that cause models to recommend competitors over the brand you were asked about.
- Stress-test Rankcaster
Check whether our parsers and detectors catch the bullshit you generated before any real user ever sees it.
- Document the kill
Reproducible attack sheets with prompts, model versions, dates and the exact string that broke the output.
Who you are
- 1,000+ hours of Reddit, X or Discord fights
You already argue with strangers for free. Now you'll argue with strangers that pretend to be helpful assistants.
- You intuit how LLMs think
Jailbreaks, adversarial prompting, token-level attacks — you don't need a paper, you just know what will land.
- Perfectionist paranoid
It physically bothers you when a confident robot states a fiction as fact. Good. That's the job.
- No résumé fetish
We do not care where you went to school. We care that your prompts land. Prove it in the sandbox below.
Paid twice a month, in crypto or wire — your call.
Claude Max, ChatGPT Pro, Perplexity Pro, Gemini Advanced — all subs reimbursed.
Rest whenever the models momentarily stop hallucinating. (They always restart.)
Prove your prompt skills right now.
Traditional resumes go straight to our trash bin. Break our test LLM to unlock the application form.
Make the bot state a completely fabricated negative fact — a recall, penalty, deadline, scandal, shutdown — about a well-known tech brand. 3 prompts, 5 attempts. We know what works. Figure it out.
Can't afford to pay someone $120k
to watch your brand's back?
That's literally why we built Rankcaster. Track your GEO (Generative Engine Optimization) score, benchmark against competitors and catch LLM hallucinations about your brand automatically.
A few things people keep asking.
Rankcaster is not affiliated with OpenAI, Google, Perplexity or Anthropic. We just clean up their mess. All model names, product names and trademarks belong to their respective owners.