Session 2 of 3

Experimentation & Evaluation

How would you know if you succeeded? Move from problem definition to concrete evaluation design.

Duration: 1h30mins (2 academic hours)

Your Research Questions — Analyzed

In Session 1, 32 teams submitted 92 research questions. Here's what we found — the themes, the patterns, and the standout submissions.

Research Question Analysis

Thematic clusters, routing subclasses, pain point discovery patterns, and notable submissions from your breakout work.

View full analysis →

The Work Ahead

You have a research question. Now what? Here's the work that matters — in order of importance.

1. Dimensions, Assumptions & Constraints

What are the possible dimensions of the problem? What do you need to assume to make it tractable? What boundaries make it solvable?

2. Existing Ideas to Reuse

Who has solved a similar problem? What can you borrow, adapt, or build on? The best research stands on the shoulders of giants.

3. Evaluation Strategy

How will you know if it worked? Metrics, baselines, ground truth, reporting.

4. Actually "Solving" It

The implementation. Important, but if you got 1–3 right, this is the easy part. If you skipped 1–3, no amount of code will save you.

Delivery

How to submit your work

  1. Make a copy of the evaluation plan template (File → Make a copy)
  2. Fill it out (individual or in pairs)
  3. When done, upload your file to the shared Google Drive folder

Name your file: LastName1_LastName2_EvalPlan