ProdigyMindset

It all starts with the right mindset. Our project, ProdigyMindset, is your personal AI partner for academic success.

Inspiration

It all started during a frantic, late-night study session in our hostel room in Bengaluru, just weeks before our end-semester exams. The floor was littered with textbooks, printed lecture slides, and half-empty cups of coffee. We were staring at a meticulously planned schedule on Google Calendar, but it felt like a work of fiction. The schedule said "Revise Data Structures," but we were stuck on AVL tree rotations for three hours, and the entire plan for the week had collapsed.

We felt like we were spending more energy managing our learning than actually learning. We were drowning in information and using static, "dumb" tools to plan a deeply dynamic process.

That was our moment of clarity. We asked ourselves: "What if our planner was more than just a calendar? What if it was an intelligent partner that actually understood our progress and adapted with us?" We were surrounded by advancements in AI, yet our study methods were stuck in the past. We were inspired to build the tool we desperately needed—a planner with a prodigy's mindset.

What it does

ProdigyMindset is an intelligent, AI-powered study planner that acts as a personal tutor, transforming the way students prepare for exams. It’s not just about scheduling; it’s about creating a dynamic and responsive learning ecosystem.

Here’s the user journey:

Personalized Plan Generation: A student inputs their subjects, uploads their syllabus or notes (PDFs, text), and sets their exam dates. ProdigyMindset uses Gemini to analyze this content and generates an optimized, prioritized study schedule.

Active Recall with AI Quizzes: After completing a study block, the app automatically generates a short, relevant quiz based on the material just covered. This reinforces learning through active recall, rather than passive reading.

Adaptive Re-planning: This is our core feature. Based on the student's quiz performance, our AI agent works in the background. If a student aces a topic, its review is scheduled further out. If they struggle, the system autonomously re-prioritizes that topic, carving out a follow-up session in the near future to address the weakness. The plan is always adapting to their real-time needs.

How we built it

We built ProdigyMindset using a modern, AI-first tech stack designed for rapid development and powerful capabilities.

Frontend: We used Next.js and TypeScript for a fast, responsive, and type-safe user interface.

AI Orchestration: The brain of our operation is Genkit, Google's new framework for building AI-powered applications. It acted as the nervous system, allowing us to define, deploy, and monitor complex AI flows with incredible speed.

Core Intelligence: We leveraged the Gemini 1.5 Pro API for its advanced reasoning and large context window.

For Planning: Gemini analyzes the syllabus and topic dependencies to create the initial smart schedule.

For Quizzes: We implemented a Retrieval-Augmented Generation (RAG) flow in Genkit. When a user uploads notes, we use Gemini to generate contextually-aware quiz questions and answers directly from the provided text.

The Agentic Loop: We designed a specific Genkit flow that triggers after every quiz submission. This flow feeds the quiz score and topic into a Gemini prompt engineered to act as a "tutor agent." The agent then decides the next best action (e.g., "Schedule a 30-minute review in 2 days") and updates the user's plan in our Firebase Firestore database, which syncs with the frontend in real-time.

Challenges we ran into

Taming the AI Planner: Our first version of the adaptive planner was too aggressive. A single failed quiz would cause the AI to frantically reshuffle the entire week, which was more stressful than helpful. We had to iterate extensively on our system prompts, building in constraints and rules to make the agent behave more like a calm, rational tutor that makes subtle, intelligent adjustments.

Ensuring Quiz Quality: Getting high-quality, non-hallucinatory quiz questions from diverse and often messy lecture notes was a major prompt engineering challenge. We spent hours refining our RAG prompts, adding few-shot examples and structured output schemas to ensure the questions were relevant, accurate, and in a consistent format.

Real-time State Sync: With the AI constantly modifying the schedule in the backend, ensuring the user's UI updated instantly and gracefully without a full page reload was a complex state management puzzle.

Accomplishments that we're proud of

We are incredibly proud of building a fully functional, closed-loop AI system. That moment when we first intentionally failed a quiz and watched the schedule intelligently re-prioritize the topic for the next day was magical. It was the proof that our core idea was not just possible, but genuinely useful. We're also proud of the on-the-fly quiz generation, which transforms any static document into an interactive learning tool. For a small team in a short amount of time, creating a system with this level of autonomy and utility is something we'll never forget.

What we learned

This project was a deep dive into the practical application of agentic AI. We learned that the future of AI products isn't just about single request-response interactions, but about creating autonomous systems that can reason and act over time to help a user achieve a goal. We also learned, firsthand, that prompt engineering is a critical skill; the quality of the AI's output is a direct reflection of the quality of the instructions it's given. Finally, we learned the importance of user experience in AI—the "magic" needs to be presented in a way that feels helpful and trustworthy, not unpredictable.

What's next for ProdigyMindset

This is just the beginning. We have a clear vision for where we want to take ProdigyMindset.

Expanded Integrations: We plan to integrate with Google Calendar, Notion, and other platforms students already use, allowing them to import tasks and export their study plans.

Deeper Analytics: We want to provide students with a rich dashboard that visualizes their learning progress, highlighting their strongest and weakest areas over time.

Collaborative Features: Imagine ProdigyMindset automatically identifying students in a class who are struggling with the same concept and suggesting they form a study group.

Institutional Version: We aim to develop a version for universities and colleges, giving educators anonymized, high-level insights into which topics their students are finding most difficult, allowing them to adjust their teaching in real-time.

Note: Running locally ensures faster inference, reduced latency, and cleaner model outputs.