AI & PromptingBeginner 2 hours

Vector Database Metadata Extraction Prompt

Extract key metadata fields from raw text as structured JSON for vector search.

The Scenario

A global recruitment agency parses thousands of resumes per day to store them in a vector database. To enable efficient hybrid search (combining semantic search with keyword/metadata filters), they need to extract key fields from raw resume text as a clean JSON payload. Crucial fields include technical skills, years of experience, current location, and salary expectations (standardized to USD).

The Brief

Design a system prompt and query structure that takes unstructured resume text as input, extracts relevant candidate information, and outputs a strict JSON object matching a predefined schema for metadata filtering.

Deliverables

  • System prompt instructions specifying the extraction rules, format guidelines, and constraint checks.
  • A JSON schema representing the target output (including fields for skills array, total experience years, location, and salary expectations in USD).
  • Three test cases showing raw input text and the corresponding JSON output generated by the prompt.
  • A fallback/safeguard strategy detailing how to handle incomplete or missing information without breaking the schema.

Submission Guidance

Format your deliverables in a single Markdown file. The prompts should clearly highlight variable placeholders and formatting constraints, and the schemas and examples should be formatted as code blocks.

Submit Your Work

Your submission is graded against the rubric on the right. If you pass, you get a public Badge URL you can share on LinkedIn. There is no draft save, so work offline first and paste your finished response here.

This appears on your public Badge.

0/20000 charactersMarkdown supported

One per line or comma separated. Up to 5 links.

Loading security check...

By submitting, you agree your submission text, name, and evaluation will appear on a public Badge URL.