Google Maestro is a research-based quiz held during the annual ACES national-level techfest. Participants race to find answers online, so every second matters and each person must receive the question at the same time.
The event previously relied on Google Forms. Admins created a new form for each question, participants repeatedly entered their details, and organizers manually sorted responses after every round. I built a dedicated platform to replace that workflow.
I was the primary developer, while other ACES members contributed to some frontend pages and testing. The application uses NestJS, React, TypeScript, and WebSockets and is open source on GitHub.
How it worked
- The organization configured its instance. Each deployment represented one organization. Its name and super-admin email were supplied at hosting time, and the initial administrator was created automatically.
- Admins prepared the event. They created a quiz, divided it into rounds, added questions, and invited participants.
- Participants joined a lobby. A unique code connected each participant to the correct quiz before the round started.
- Questions were delivered in real time. WebSockets pushed the same question to every connected participant together. Answers were timestamped, and paste was disabled to support the competition’s rules.
- Admins verified answers. When everyone responded or the timer ended, an administrator reviewed the submitted answers and marked the correct ones.
- Scores and rankings updated. The platform accumulated scores across questions and rounds before producing the final leaderboard.
The main engineering challenge
The difficult part was coordinating state across two different views: the administrator controlling the quiz and participants answering it. Starting a question, closing submissions, reviewing answers, updating scores, and moving forward all needed to happen in a predictable order.
The WebSocket layer made question delivery immediate, but it also required clear event boundaries and server-owned state. The server determined which question was active, whether answers were still accepted, and when the quiz could advance. That prevented clients from deciding important competition state independently.
Accurate timestamps were also important. In a speed-based event, submission order had to come from the server rather than each participant’s device clock.
Result
Google Maestro was used during the ACES techfest with around 20 participants across five rounds. It removed repeated form creation and participant registration from the middle of the event while giving organizers one place to control questions, verify answers, and track scores.
The project is intentionally focused on ACES’s event workflow rather than being a generic quiz product. Its public repository documents the implementation and makes the work directly verifiable.