AI Prompt Engineer Needed to Build Classification Prompt for Industry Identification in Clay.com

I’m looking for an experienced AI prompt engineer to help create a highly accurate prompt (or series of linked prompts) that classifies which industries a company serves based solely on its website and LinkedIn content. The output must be extremely reliable — this prompt will be used in a production setting and accuracy is essential. The goal is to build a system that, when given a company name and website, can determine whether the organization serves one or more of the following markets:• K–12 Education• Higher Education (colleges and universities)• Corporate Learning / Workforce Development• Other / Not Education (should specify the industry)Some companies serve multiple industries, and the prompt must be able to output all relevant markets.What I Need Built:• A single prompt or structured prompt chain that instructs the model to scan a company’s website (and optionally LinkedIn) for explicit evidence of the industries it serves. • A clear set of classification rules that prevents false positives — especially around ambiguous terms like “students,” “learning,” or “education” that do not necessarily imply K–12 or higher ed. • Strong logic that keeps the model from inferring or guessing. The AI should rely only on explicit evidence. • The ability for the prompt to stop scanning once it finds qualifying evidence for an industry.• A fallback process where, if the model finds no evidence of serving K–12, Higher Ed, or Corporate Learning, it returns the correct alternative industry (healthcare, finance, nonprofit, etc.). • Explicit disqualification of membership organizations, networking councils, associations, chambers, community groups, etc., unless they clearly sell products or services to education institutions. Required Output Format:The AI should return:• One or more industries served: K–12, Higher Education, Corporate Learning, or Other (with industry name).• A short explanation describing where it found the evidence. • The exact page URL where the evidence was located. • If no evidence is found: “Other – Industry Not Education,” with reasoning. Ideal Skills:• Proven experience designing prompts for GPT-4/4.1, and Claygent• Experience with LLM reasoning optimization and pattern-based classification• Understanding of website structure, industry language, and customer-segment cues• Ability to write prompts that enforce strict logic and avoid hallucinations• Optional bonus: experience designing prompt chains or multi-step evaluation flowsAboutThis Project:I’m building a lead-qualification workflow that depends heavily on the AI’s ability to accurately identify whether a company sells into the K–12, Higher Education, or Corporate Learning markets.Getting this right is mission-critical. I’m looking for someone who can help engineer a solution that’s accurate, repeatable, and stable across thousands of evaluations. If you can deliver a prompt or prompt system that consistently produces correct classifications, I’d love to work with you. Apply tot his job