Inspiration

This idea struck me late at night, just after I took a solid piss at 3:29 a.m. My back hurts, my knees are creaking, and my ankles are just about to snap. Then I am reminded of the countless sleepless nights that I’ve endured as a failing student while tutoring others so I could keep my head above the water for the past decade. “Blimey, I wish there was a cheeky shortcut to proper teaching and learning, innit?”— And the rest is history.

// Here are some statistics so that I can prove my point x)

Research shows teacher burnout impacts 44–53% of educators, with attrition rates doubling to 9% and a projected 44 million shortage by 2030 due to workloads and poor balance on a global level.
In Hong Kong, 87.6% of teachers report high stress, and over 6,700 resignations in 2022–2023.

Hong Kong students often face time management challenges, with 50% of secondary students using over half their available time spontaneously without planning during COVID-19 school suspensions. This contributes to academic burnout, high stress levels, and sleep insufficiency.

Financial barriers worsen inequities, including a 27% child poverty rate, where 45.8% of low-income students believe finances limit their learning opportunities. Low-income undergrad enrollment remains below 15%, and ethnic minorities suffer high dropout rates due to inadequate resources and subsidies.

What it does

Our platform aims to assist teachers in efficiently managing their courses, materials, assignments, and grading. A key feature is automated progress reports, which provide teachers with a clear overview of each student's performance without any manual effort, enabling better support.

At the same time, students receive personalised learning experiences through:

  • AI-generated practice questions
  • Automated progress reports
  • Predicted grades
  • Editable weekly timetables

To encourage students’ engagement, we designed reward systems that students may find appealing or motivating, such as:

  • A customizable profile
  • A customizable personal v-pet

Challenges we ran into

One of the challenges we ran into while speedrunning this abomination was hitting the rate limit for AWS Bedrock model requests — my request frequency was too high, and I had to slow down and check the Bedrock model invocation quotas to stay within the allowed limits. There was also a bit of miscommunication, but it really didn’t affect our performance too much.

Accomplishments that we're proud of

Honestly, considering the sleep deprivation, caffeine overdose, and the occasional existential crisis, I’d say we did a bloody good job. We managed to turn a crazy midnight idea into a fully functional prototype that actually works — and dare I say, works quite decently.

From building an AI-backed progress tracker to making the v-pet system come to life (it blinks now!), our team managed to mash together every component we dreamt up in a matter of hours. The fact that we didn’t crash AWS permanently or set our laptops on fire feels like an even bigger win.

What we learned

Apart from realising that time management isn’t just for students, we also learned more about teamwork than any group project ever taught us. AWS threw us curveballs (rate limits, authentication headaches, you name it), but we came out knowing how to wrestle APIs, deploy models, and debug under pressure — sometimes literally three minutes before the deadline.

Most importantly, we learned how to balance creativity with practicality: that not every ridiculous idea can be implemented in 24 hours… but maybe that’s the beauty of it.

What's next for Omniscient

Omniscient means all knowing. It doesn’t end here — this is just our first chapter. We want to evolve beyond being another “academic helper” app and become an integrated ecosystem where teachers teach smarter, and students learn better.

We’re planning to incorporate:

  • Adaptive learning engines
  • Group analytics for classes
  • Community-driven resources

Built With

Share this project:

Updates

Submission history