DMP Chef (Beta)
Overview
DMP Chef is developed in collaboration with DMPTool.org as a testing ground for AI-assisted Data Management Plan (DMP) drafting. Its purpose is to experiment, iterate, and gather community feedback on AI-based pipelines that could help researchers create funder-compliant DMPs with less effort. Validated and production-ready capabilities developed here will ultimately be incorporated into DMPTool.org .
DMP Chef is developed in collaboration with the California Digital Library, through a grant from the University of California Office of The President (UCOP) and the Chan Zuckerberg Initiative (CZI).
Key Features
AI-Assisted Drafting
Generate funder-compliant Data Management Plans (DMPs) automatically using AI models trained on real funder templates and FAIR data principles.
Multi-Funder Support
Create DMPs tailored to specific agencies including NIH, NSF, and other national and international funding bodies.
Export Options
Easily export finalized DMPs in DOCX, PDF, or machine-actionable JSON formats for seamless submission and archiving.
Snapshots
Core Technologies
Powering DMP Chef with advanced AI, retrieval, and data optimization.
Prompt Engineering
Custom-designed prompts ensure precise, context-aware responses that align with funder requirements and research domains.
Retrieval-Augmented Generation (RAG)
Combines AI generation with real-time access to structured knowledge bases for accurate, up-to-date content.
Fine-Tuned Language Models
Models optimized on research data policy language and DMP examples to improve factuality and style consistency.
Vector Databases
High-performance semantic search supports retrieval of relevant policy documents and prior DMPs.
Open-Source Infrastructure
Built with transparency and extensibility in mind, enabling community-driven development and easy institutional adoption.
Project Roadmap
Ready to revolutionize your Data Management Plans?
Draft, refine, and share funder-compliant DMPs with ease.