NED3 Open Engineering AI
Two open resources support different stages of responsible AI-enabled engineering work. They share a physics-first, evidence-aware approach, but neither replaces engineering judgment, source verification, or responsible authorship.
| I want to… | Resource | Start with |
|---|---|---|
| Learn or teach AI/ML for physical systems | Machine Learning for Engineers | The course gateway, First 30 minutes, or adoption guide |
| Apply AI agents to an active engineering research task | Mechanical Engineering Research Skill | Its quickstart, beginning with a task the researcher can independently inspect and verify |
How They Fit Together
The curriculum develops the habits needed to formulate a physical problem, establish a baseline, evaluate a model, inspect failure cases, and communicate a defensible conclusion. The research-workflow plugin helps researchers apply related habits to literature synthesis, proposal development, data analysis, technical writing, presentations, and mentoring.
Use the curriculum when the goal is learning or instructional adoption. Use the plugin when the goal is a bounded research workflow that a researcher can independently inspect and verify. Both are open resources maintained through their canonical GitHub repositories.
Responsible Use
- Do not treat agent output as evidence without checking the underlying sources, data, model, and physical interpretation.
- Do not upload confidential proposals, student records, sponsor-restricted data, unpublished intellectual property, or credentials to public tools or issue trackers.
- Preserve data provenance, units, assumptions, evaluation conditions, and limitations when adapting materials or reporting results.
- Cite the applicable resource and identify the version or commit used when it materially supports a scholarly or public workflow.
Get Involved
Report reproducibility problems, unclear instructions, and reusable improvements through the repository issue trackers or contribution guides. When publicly adapting course material or a workflow, a concise description of the setting, evidence, and improvement helps other users judge whether it transfers to their work.