HelixCipher/job-training-data

1---2language: en3pretty_name: Job Training Data for JSON Extraction4task_categories:5- text-generation6tags:7- json8- web-scraping9- job-postings10- training-data11- json-extraction12size_categories: medium13---14 15# Job Training Data16 17Training dataset for fine-tuning LLMs to extract structured JSON from job postings.18 19## Description20 21This dataset contains **12,000 examples** of job postings in markdown format paired with their JSON extractions. Used to train the `job-posting-extractor-qwen` model.22 23## Data Format24 25Each example contains:26 27- `instruction`: What to do (e.g., "Extract job fields as JSON").28 29- `input`: Job posting in markdown format.30 31- `output`: Expected JSON output (as a string).32 33### Example Entry34 35```json36{37  "instruction": "Extract all job fields as JSON object.",38  "input": "# Job Position\n**Position:** Platform Engineer\n**Company:** DATAECONOMY\n**Location:** Charlotte, NC\n\n## Job Description\nRole: Platform engineer...",39  "output": "{\"job_title\": \"Platform Engineer\", \"company\": \"DATAECONOMY\", \"location\": \"Charlotte, NC\"}"40}41```42 43## How It Was Created44 451. Data sourced from webscraped job postings.46 472. Converted to markdown using template-based generation.48 493. JSON labels programmatically extracted from the scraped data.50 514. Augmented with 15 instruction variations.52 53## Dataset Statistics54 55| Metric | Value |56|--------|-------|57| Total examples | 12,000 |58| Unique JSON fields | 7 (job_title, company, location, work_type, description, experience_level, salary) |59| Instruction variations | 15 |60 61## Files62 63- `job_training_data.json`: Main training data (12,000 examples).64 65## License & Attribution66 67This dataset is licensed under the **Creative Commons Attribution 4.0 International (CC BY 4.0)** license.68 69You are free to **use, share, copy, modify, and redistribute** this material for any purpose (including commercial use), **provided that proper attribution is given**.70 71### Attribution requirements72 73Any reuse, redistribution, or derivative work **must** include:74 751. **The creator's name**: `HelixCipher`76 772. **A link to the original repository**:  78 79   https://github.com/HelixCipher/fine-tuning-an-local-llm-for-web-scraping80 813. **An indication of whether changes were made**82 834. **A reference to the license (CC BY 4.0)**84 85#### Example Attribution86 87> This work is based on *Fine-Tuning An Local LLM for Web Scraping* by `HelixCipher`.  88> Original source: https://github.com/HelixCipher/fine-tuning-an-local-llm-for-web-scraping89 90> Licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0).91 92You may place this attribution in a README, documentation, credits section, or other visible location appropriate to the medium.93 94Full license text: https://creativecommons.org/licenses/by/4.0/