StudyPilot AI dashboard for uploading study materials, AI tutoring, practice quizzes, and tracking learning progress.
Quick actions enable material uploads, AI tutoring, and practice quizzes. Recent Study Materials displays uploaded documents.
Upload PDF documents or paste study notes to extract text with page numbers for AI tutoring and quiz generation.
Uploaded a PDF study document successfully. StudyPilot AI extracted 307 words from one page and set it as the active study material.
AI Tutor provides source-based answers with page citations and flags insufficient evidence when needed.
Practice Quiz generates multiple-choice questions from study materials to test knowledge, with protected answers and instant feedback.
Practice quiz results show a 33.3% score, with detailed answer reviews and page references to help identify learning gaps.
Learning Summary tracks quiz scores, highlights weak concepts, and recommends pages to review with AI Tutor for targeted improvement.

Inspiration

Students often struggle to learn effectively because their study materials are scattered across lecture videos, textbooks, and presentation slides. Finding relevant information, identifying weak topics, and preparing for exams can take significant time. General-purpose AI chatbots may also provide answers that are not grounded in a student's actual course materials.

We created StudyPilot AI – Personalized Learning Companion to make learning more organized, interactive, and personalized. Our vision is to help students understand concepts, practice what they learn, and focus on the topics where they need the most improvement.

What it does

StudyPilot AI is an AI-powered learning companion designed to transform course materials into a personalized study experience.

Key features include:

Multimodal Learning: Process educational materials such as textbooks, lecture slides, and lecture videos.
Source-Grounded AI Tutor: Answer student questions using uploaded learning materials, with citations pointing to relevant pages, slides, or video timestamps.
Adaptive Assessments: Generate multiple-choice questions, short-answer questions, and practice tests based on selected topics and difficulty levels.
Personalized Feedback: Explain answers, identify mistakes, and highlight topics that need further revision.
Learner Progress Tracking: Maintain topic-wise mastery estimates and use assessment results to recommend what students should study next.
Reliable Responses: Identify questions that are not covered by the uploaded materials and clearly communicate when sufficient supporting information is unavailable.

Our goal is to make studying more efficient by connecting learning resources, assessments, and personalized recommendations in one place.

How we built it

StudyPilot AI is designed around a modular architecture that combines document processing, information retrieval, generative AI, assessment generation, and learner modeling.

The system processes educational resources and organizes their content into searchable knowledge units. A retrieval-augmented generation (RAG) pipeline retrieves relevant content before generating answers, helping connect responses to their original sources.

The tutoring component uses retrieved evidence to answer questions and provide source references. The assessment component generates practice questions and feedback from the selected material. The learner model uses assessment performance to estimate topic-wise mastery and guide subsequent learning recommendations.

The application brings these capabilities together through a web-based interface for uploading materials, interacting with the tutor, taking assessments, and reviewing learning progress.

Challenges we ran into

One major challenge is maintaining accurate connections between generated answers and their original learning materials. Source citations must identify relevant content rather than simply provide general references.

Another challenge is extracting useful information from different formats, especially diagrams, presentation slides, and lecture videos. Converting these resources into a consistent, searchable knowledge base requires careful processing.

Generating reliable questions and answer keys is also challenging. Incorrect assessments can lead to misleading mastery estimates and ineffective recommendations.

Finally, personalization requires more than generating different questions. The system must use student performance to update its understanding of learning needs and demonstrate meaningful changes in subsequent recommendations.

Accomplishments that we're proud of

We are proud of the vision of bringing learning resources, AI-powered tutoring, assessments, and personalized study recommendations into a unified learning companion.

StudyPilot AI focuses on three important principles: grounding answers in learning materials, making assessment feedback useful, and adapting study recommendations to individual learning needs.

By combining retrieval-based knowledge access with assessment-driven personalization, the project aims to provide a more trustworthy and structured alternative to studying with disconnected resources and general-purpose chatbots.

What we learned

This project helped us understand that building an effective educational AI system requires more than connecting a language model to a chatbot interface.

Reliable retrieval, accurate citations, question verification, learner modeling, and systematic evaluation are essential to the learning experience. We also learned the importance of separating source-supported information from unsupported answers and measuring system performance rather than relying only on impressive demonstrations.

We gained a deeper understanding of how generative AI, information retrieval, multimodal processing, and adaptive learning techniques can work together to support personalized education.

What's next for StudyPilot AI – Personalized Learning Companion

Our next steps include improving multimodal extraction, strengthening citation accuracy, and introducing more reliable question verification.

We aim to develop a more effective learner model that tracks mastery across sessions and recommends revision based on weak topics and forgetting patterns. We also plan to evaluate retrieval quality, answer faithfulness, assessment correctness, and question repetition using a structured benchmark.

Future enhancements may include multilingual tutoring, audio-based learning sessions, visual course maps showing topic prerequisites, personalized revision materials, and study schedules based on upcoming examinations.

Our long-term vision is to make StudyPilot AI a dependable learning companion that helps students learn at their own pace, understand their progress, and make better use of their study time.