Best Practices
This guide outlines recommended practices for achieving optimal performance when working with the library.
Getting Started with the Library
For New Users (Students)
typescript
// DO: Start with a well-defined problem
const project = await daniel.mentor({
level: 'student',
approach: 'guided',
scope: 'well-defined',
meetingFrequency: 'weekly'
});
// DON'T: Try to solve everything at once
const project = await daniel.mentor({
scope: 'boil-the-ocean' // SCOPE_OVERFLOW error
});Key principles:
For Experienced Users (Collaborators)
typescript
const collaboration = await daniel.collaborate({
mode: 'peer-to-peer',
communication: 'async-first',
codeReview: true,
whiteboard: 'on-demand'
});Research Workflow
The Ideal Pipeline
- Start with the physics — What question are you actually trying to answer?
- Literature review — What has been tried? (Check arXiv, not just Google)
- Prototype fast — Get a minimal version working before optimizing
- Measure everything — If you can't plot it, you don't understand it
- Write as you go — The paper is not something you do "at the end"
Common Anti-Patterns
typescript
// Anti-pattern: Premature optimization
const model = new GNN({ layers: 47, features: 2048 });
// Start simple. Add complexity only when the simple version fails.
// Anti-pattern: Training without understanding
model.fit(data); // "It works!" — but do you know WHY?
// Best practice: Understand, then scale
const baseline = new SimpleModel();
const results = await baseline.evaluate(data);
// Now you have something to compare against.Communication Protocols
| Method | Best For | Response SLA |
|---|---|---|
| Formal requests, paper drafts | ||
| Slack/Mattermost | Quick questions, links | |
| In-person | Whiteboarding, debugging, brainstorming | |
| Video call | Remote collaboration |
Supervision Style
typescript
const supervisionConfig = {
style: 'collaborative',
meetings: {
frequency: 'weekly',
duration: '1h',
format: 'discussion-not-status-update'
},
expectations: {
independence: 'high',
communication: 'proactive',
mistakes: 'encouraged-and-expected'
}
};