striver

Inspiration

Learning is not the same for everyone. Some students understand concepts better through visuals, while others prefer explanations, examples, practice questions, or simply having someone explain a difficult topic step by step. Traditional learning systems often provide the same content to everyone, regardless of their learning pace or difficulties.

We were inspired by the idea of creating a learning companion that could actually adapt to the student, rather than expecting the student to adapt to the learning platform.

That led us to build Striver, a multimodal AI-powered personalized tutor designed to make learning more interactive, adaptive, and accessible.

What it does

triver acts as a personalized AI tutor that helps students understand concepts based on their individual learning needs.

Instead of simply giving the same answer to everyone, Striver aims to adapt explanations based on the learner's interaction, level of understanding, and preferred way of learning.

The system can support different forms of learning such as:

Text-based explanations for step-by-step understanding Visual learning to make complex concepts easier to grasp Interactive questioning to check understanding Personalized explanations based on the learner's level Practice and feedback to identify areas that need improvement Adaptive learning that adjusts the learning experience as the student progresses

Our goal is to make the experience feel less like searching for an answer and more like having a tutor who understands how you learn.

How we built it

We started by identifying the core problem: students often get stuck because learning resources are not personalized to their current level or learning style.

We designed Striver around a multimodal AI approach, combining user interaction with AI-generated explanations and learning assistance.

The development process involved:

Defining the learner journey and major use cases. Designing the interface around a simple and distraction-free learning experience. Building the core application and connecting it with AI capabilities. Designing prompts and logic to generate explanations appropriate to the learner's context. Adding adaptive interactions so the system can respond differently depending on the learner's progress. Testing different learning scenarios and refining the responses and user experience.

We focused not only on generating answers, but also on making the AI explain concepts clearly and support the learning process.

Challenges we ran into

One of our biggest challenges was making the AI behave like a tutor rather than simply an answer generator.

A correct answer is not always a good learning experience. Sometimes a student needs a simpler explanation, an analogy, another example, or a hint instead of the complete solution.

We also faced challenges around:

Designing effective adaptive learning interactions Keeping AI responses clear and understandable for different learner levels Handling multimodal inputs consistently Reducing the possibility of incorrect or misleading AI-generated information Balancing personalization with a simple user interface Testing different learning scenarios within the limited time available during the hackathon

These challenges helped us understand that building an educational AI system requires more than connecting an AI model to an interface—it requires thinking carefully about how students actually learn.

Accomplishments that we're proud of

We are proud of turning the idea of a personalized AI tutor into a working prototype within the hackathon timeframe.

Some aspects we are particularly proud of are:

Building a multimodal learning experience Creating an AI tutor focused on personalization rather than generic responses Designing interactions around the learner's understanding and progress Making the interface simple enough for students to use without a steep learning curve Exploring how AI can be used to provide more accessible and individualized education Successfully bringing together the AI, application logic, and user experience into one project

Most importantly, Striver helped us explore how technology can move from "giving students answers" to "helping students learn."

What we learned

this project taught us that personalization is one of the most important parts of building AI for education.

We learned how important it is to:

Design AI interactions around real user needs Write prompts that encourage explanations instead of just answers Think about different learning levels and preferences Handle multimodal information effectively Test AI systems with different types of users and learning scenarios Consider reliability and responsible AI when building educational tools Iterate quickly based on what works and what doesn't

We also learned that a good AI tutor should know when to explain, ask, guide, give a hint, or encourage the learner rather than always providing the final answer.

What's next for striver

we see Striver as the beginning of a larger personalized learning platform.

Our next steps include:

Improving learner profiling and personalization Building stronger adaptive learning paths Adding more multimodal learning capabilities Introducing personalized quizzes and assessments Tracking learning progress and identifying knowledge gaps Supporting more subjects and academic levels Improving accessibility through voice and language support Making explanations more reliable through stronger grounding and verification Exploring teacher/mentor dashboards to monitor student progress

Ultimately, we want Striver to become an AI learning companion that grows with the learner—helping students understand what they are learning, where they are struggling, and how they can improve, one step at a time.

Built With

• adaptive-learning
• ai
• ai-tutor
• computer-vision
• education-technology
• generative-ai
• machine-learning
• multimodal-ai
• natural-language-processing
• personalized-learning
• voice-ai
• webapplication