Brainoe

Inspiration

My friend recently shared a Claude artifact of a custom quiz app he built to study for an upcoming test. I was blown away by both the simplicity of the concept and the execution speed, which inspired me to build my own.

Without an active Claude subscription, I initially attempted to recreate it using Canvas in Gemini Enterprise. While the output was decent, the workflow took considerably longer (~3 hours in Gemini vs. <30 minutes in Claude). Since "decent" wasn't my goal and university coursework was piling up, I shelved the project until Winter Break.

When I signed up for this Hackathon, I realized it was the perfect opportunity to revive the project—allowing me to build the tool while gaining hands-on mentorship on how to orchestrate AI coding agents (effectively killing two birds with one stone).

What It Does

In its current build, Brainoe includes the following core capabilities:

Interactive Quiz Runner: Delivers questions dynamically from a central database and provides instant feedback upon submission.
Scoring & Summary: Calculates and presents an end-of-quiz score breakdown.
Persistent History (SQLite): Replaced temporary JSON storage with a local SQLite database, ensuring quiz attempts and settings persist across restarts.
Progress Analytics (Overview Dashboard): Visualizes performance metrics using stored quiz logs, including:
- Day Streak
- Daily & Weekly Accuracy
- History Calendar view
Question Bank Hub: Supports manual creation, editing, and deletion of questions, alongside an experimental PDF importer (currently Python-based).

How I Built It

VS Code Agents & GitHub Copilot: Leveraged AI coding agents directly inside VS Code to write, debug, and refactor code.
Agentic Workflows & Skill Files: Followed structured prompts and specialized skill guidelines curated by the hackathon organizers to keep agent outputs aligned with project architecture.

Challenges I Ran Into

My biggest hurdle was getting fixated on building a fully polished, final product rather than prioritizing a lean Proof of Concept (PoC).

The coding agent repeatedly had to steer me away from overly ambitious features to keep the scope aligned with the hackathon criteria. In hindsight, starting with a lightweight core was the right architectural decision. To prevent losing good concepts without derailing the build, I had the agent document future features in a dedicated handover.md file. This preserved the broader product roadmap while keeping the active codebase stable.

Accomplishments I'm Proud Of

Solid Foundation: Established a modular architecture ready for feature expansion.
Agentic Mastery: Learned how to effectively steer and collaborate with AI coding agents.
Working PoC: Delivered a stable, fully functional prototype within the competition window.
Insightful Brainstorming: Used conversational agent interactions as a sounding board to clarify the app's long-term feature set.

What I Learned

How to configure, prompt, and guide AI coding agents inside modern IDEs.
Practical Git & GitHub repository workflows, including commit discipline and version control history.
Core full-stack application architecture and state lifecycle patterns.

What's Next for Brainoe

Enhanced Document Parsing & OCR
- Upgrade the ingestion pipeline with OCR and specialized agents to handle multi-column exam formats, detached answer keys, and scanned test papers.
Deduplication Engine
- Add intelligent diffing and merge/replace prompts during imports to avoid redundant questions.
Multi-Module & Subject Organization
- Enable dedicated database partitioning per academic course or topic.
Targeted Quizzing & Tagging
- Introduce granular topic tags for custom, modular study sessions.
AI Question Synthesis & Note Ingestion
- Generate high-yield practice variations (e.g. dynamic numerical values, alternative MCQ distractors) and parse raw lecture slides directly into question banks.
Weak-Topic Diagnostics & Timed Modes
- Build an in-depth Analytics Page identifying knowledge gaps, paired with timed simulations for exam readiness.
Immutable Review Snapshots
- Lock historical test review states so retroactively editing bank questions doesn't alter past review accuracy.
Cloud Sync & Cross-Device Access
- Introduce user authentication and cloud synchronization for seamless mobile and desktop study sessions.