Team:

• Khang Hou, a final-year Computer Science undergraduate passionate about bridging the gap between technology and our everyday lives

• Arjun Kumar, currently founding engineer at Smithery AI. Graduated from NUS in ‘22.
  Works on projects outside of work, intersection of education and AI being one of them

• Soh Enxi, Destiny, a 1st year Architecture undergraduate who is passionate about bridging the gaps in the social aspect of society through serving the community

• Chloe Choo, a first year business and media comms student who likes to work on projects in various sectors for marketing and media, including but not limited to education and innovative technology

​

Project Overview:

smartbeans is a digital ecosystem designed to bridge the gap between tuition sessions and independent study for PSLE Math students. By providing a seamless interface for targeted pre-learning, we empower tutees to master the MOE syllabus through data-driven practice. The platform doesn't just benefit the student; it creates a transparent feedback loop by notifying parents and tutors of real-time progress, specifically highlighting mastery levels and persistent "red-flag" weakness areas.

 

Why the team applied for CCSGP Fellowship:

Our team is driven by a commitment to educational equity. smartbeans began as a conceptual draft, but the CCSGP Fellowship provided the essential resources required to pivot from a prototype to a polished, user-centric product. The fellowship allowed us to refine our vision and develop a tool capable of supporting students across all socioeconomic backgrounds.

 

About the Project:

smartbeans acts as a centralised intelligence hub for tutors and agencies. By manually tagging and organising our database based on the PSLE Math MOE syllabus, we eliminate the guesswork in identifying learning gaps.

​

• Automated Question Generation Pipeline: We have removed the technical barrier for educators. Using an "Interface-First" approach, admins can generate engaging questions related to their favourite games or shows, that are more importantly syllabus-aligned
  MCQs using SOTA LLMs without writing a single line of code.

• The "Protege" Effect (Learn through Teaching): We’ve leveraged AI to unlock a high-level pedagogical technique: learning by teaching. In this mode, an LLM acts as a "protege" that provides a flawed explanation. The student must "teach" the AI by
  identifying and correcting the logic, shifting them from passive consumers to active
  instructors.

​

What the team learned from the CCSGP experience:

Building smartbeans was an intensive masterclass in full-stack product management.

​

Beyond the code, we gained vital professional competencies:

​

1. High-Pressure Delivery: Mastering the art of shipping functional features within tight fellowship milestones.
2. Scope & Expectation Management: Learning to define a "Minimum Viable Product" (MVP) that provides value without succumbing to feature creep.
3. Velocity vs. Quality: Navigating the critical trade-offs between rapid development and maintaining a stable, high-quality codebase.
4. User-Centric Optimisation: We learned that "feel" is a feature. To ensure a great first impression, we implemented severe performance optimisations, including server-side streaming and perceived-performance indicators (loading states) to minimise friction
for the end-user.

​

Through this journey, we didn't just build an app; we built a vision. We aim for smartbeans to be the definitive first thought for any parent or tutor preparing a child for the PSLE.

​

smartbeans: Making learning fun.

​

Screenshots:

​