| Field | Value |
|---|---|
| Proposal ID | LAB- |
| Title | |
| Author | |
| Date | |
| Status | Draft / Submitted / Approved / Rejected |
<2-3 sentences describing the problem this addresses>
Who has this problem?
How big is the problem? <Quantify if possible: users affected, cost, time wasted>
<2-3 sentences describing what you want to build/explore>
Why might this work? <Brief rationale - prior art, technical feasibility, intuition>
We believe that [action/solution] will result in [outcome] for [target user/system]. We will know this is true when [measurable signal].
| Metric | Baseline | Target | Measurement |
|---|---|---|---|
| <Metric 1> | |||
| <Metric 2> |
Stop the experiment if:
- <Condition 1 that means we should stop>
- <Condition 2 that means we should stop>
- Time box exceeded without meeting progress milestones
| Resource | Amount | Duration |
|---|---|---|
| Engineers | <X FTE or % time> | |
| Compute | <GPU hours, instances, etc.> | |
| Data | N/A | |
| Budget | <$X for tools, APIs, etc.> | One-time / Monthly |
Total time box: (hard stop)
| Risk | Likelihood | Mitigation |
|---|---|---|
| Low/Med/High | ||
| Low/Med/High | ||
| Low/Med/High |
If successful, how does this become a product?
| Phase | Description | Owner |
|---|---|---|
| Experiment | This proposal | Lab team |
| Evaluation | Formal eval against criteria | Lab + stakeholders |
| Handoff | Transfer to product team | Lab → Product |
| Productization | Build production version | Product team |
Estimated time to production (if successful):
| Alternative | Why Not First Choice |
|---|---|
| <Alternative 1> | |
| <Alternative 2> | |
| Do nothing |
| Role | Name | Decision | Date |
|---|---|---|---|
| Lab Lead | Approved / Rejected | ||
| Resource Owner | Resources confirmed |
- Create experiment card
- Set up tracking and measurement
- Begin experiment on
- First checkpoint on