← All articles
Learning notes6 min read

Hackvember #3: Beyond the Finish Line

How a multiplayer racing game evolved into an AI-assisted learning experience for Indonesian teachers.

Over the past few weeks, I've been exploring how Hackvember could provide value to both students and teachers. My initial focus was on creating an engaging multiplayer racing game, followed by simplifying how teachers prepare learning activities using existing materials.

But I started thinking about another part of the experience that I had barely considered: what happens after the game ends?

In most competitive games, the finish line determines the winner. Students celebrate their positions, review their scores, and move on. That's perfectly fine for entertainment, but I wanted to explore whether the information collected during the game could also support teachers in their next lessons.

That question eventually led to two new product concepts: Race Debrief and Pit Crew.

A Leaderboard Doesn't Tell the Whole Story

Imagine a classroom with 30 students participating in a race containing ten educational questions. At the end, the application displays the leaderboard, showing who finished first and how many questions each student answered correctly.

From a student's perspective, that might be enough. But for teachers, I think there's an opportunity to provide more meaningful information.

During a session, the application could collect data such as selected answers, correctness, and response times. If 18 out of 30 students answered the same question incorrectly, that might indicate a concept worth reviewing. Similarly, if several students repeatedly struggled with questions related to the same topic, the teacher might want to investigate further.

Response time is another interesting metric, although I wouldn't treat it as a direct measurement of student ability. A student who takes longer might be carefully reasoning through a question, while someone who responds immediately might be guessing. Question difficulty, reading time, and distractions can also affect how quickly students respond.

Rather than automatically labeling students based on their performance, I want the platform to present evidence that teachers can interpret using their professional judgment.

This became the foundation of Race Debrief.

Race Debrief: Learning Insights Through the Racetrack

Existing educational platforms already provide performance analytics. Wayground, for example, offers detailed assessment reports, including question-level performance and student responses. Simply building another analytics dashboard wouldn't provide much differentiation.

I began imagining something more closely connected to the racing experience itself.

What if teachers could revisit the racetrack after the session and see where students encountered difficulties? Each checkpoint could represent a question, with visual indicators highlighting those that students frequently answered incorrectly.

For example, checkpoint four might show that 60% of students selected incorrect answers. The teacher could open that checkpoint to review the question, answer distribution, and associated learning topic.

Instead of separating gameplay and assessment into completely different experiences, the racetrack could serve both purposes. During the game, it represents competition and progress. After the game, it becomes a way for teachers to explore learning evidence.

I liked this direction because it connects the product's visual identity with its educational purpose. However, I also realized that identifying difficult questions is only part of the problem.

Teachers still need to decide what to do with that information.

Pit Crew: An AI Teaching Partner

This led me to another idea: an AI assistant that teachers could interact with after reviewing their race results.

I decided to call it Pit Crew, inspired by the team that supports a racing driver during a competition.

The concept is to make AI part of the post-race evaluation process. Instead of simply displaying automatically generated recommendations, the application would allow teachers to discuss the results, explore possible explanations, and develop follow-up activities.

For example, imagine that most students answered a question about fraction division incorrectly. A teacher might ask Pit Crew why that happened or whether the topic should be explained again.

The AI could examine the question, answer distribution, and related learning objectives before suggesting possible explanations. But I also want teachers to be able to challenge those suggestions.

Perhaps the teacher noticed that students were confused by the wording rather than the mathematical concept itself. They could explain that observation to Pit Crew, and the AI should reconsider its recommendation instead of treating its original interpretation as a fact.

This collaborative interaction is what interests me most. I don't want AI to replace a teacher's judgment. I want it to help teachers make better-informed decisions.

Making Pit Crew Understand Indonesian Education

Another idea emerged while exploring the AI assistant: what if Pit Crew understood the Indonesian education system, particularly the conditions of public schools?

A generic AI might suggest reasonable teaching activities without considering the teacher's actual classroom. For example, it could recommend a digital presentation or an activity requiring every student to have a device, even when those resources aren't available.

I want to investigate whether Pit Crew could consider the teacher's curriculum, grade level, learning objectives, available equipment, and teaching time before recommending a strategy.

Indonesia already publishes official educational guidance that could support this approach. The Ministry's Panduan Pembelajaran dan Asesmen 2025, for example, emphasizes using assessment evidence to adjust future instruction. There is also official guidance on using AI to support teachers while preserving their professional judgment.

These documents could become part of a curated knowledge base that helps ground Pit Crew's recommendations in Indonesian educational practices.

Imagine a teacher asking, "I have ten minutes tomorrow, 35 students, and only a whiteboard. How should I review this topic?"

Instead of proposing an elaborate activity, Pit Crew could suggest a short explanation, a few practice questions, and a simple way to check students' understanding.

Of course, this isn't automatically unique. Other educational platforms already offer AI-powered analytics and teaching assistance. The potential differentiation would come from connecting actual classroom results, Indonesian educational context, and conversational decision-making into a simple experience.

There are also important challenges involving AI accuracy, student privacy, and the risk of drawing incorrect conclusions from limited assessment data. Those problems will need careful consideration as I move toward implementation.

From a Racing Game to a Learning Loop

Looking back, the concept has evolved significantly from what I imagined during the first day of brainstorming.

I started with a multiplayer racing game where students answered questions to reach the finish line. Now, I'm exploring three connected experiences: Race Builder, which helps teachers prepare questions from existing materials; Race Experience, where students compete through interactive learning; and Race Debrief with Pit Crew, where teachers review the results and discuss possible next steps.

Eventually, I want to connect these experiences into a continuous learning cycle. Teachers could identify a difficult topic, prepare a follow-up activity, and then observe whether students demonstrate better understanding in another session.

The idea is to connect preparation, engagement, evaluation, and teaching improvement rather than treating them as separate activities.

I still don't know whether this approach will provide enough value for teachers to adopt it. The concept needs prototypes, testing, and eventually feedback from real users. For now, it represents a product direction rather than completed functionality.

But this exploration has changed how I think about Hackvember. The 3D racing game might be what makes students excited to participate, while the experience after the race could be what makes teachers want to return.

The race might end at the finish line, but the learning shouldn't.


References