AI Homework Helper

Objective:

Develop a prototype application that leverages the Gemini API to assist students with homework by delivering both the final answer and a detailed, step-by-step explanation.

How it Works:

Students can either type their question or upload a photo of a handwritten or printed problem. The app utilizes Gemini AI to generate:

1. The final answer
2. A clear, easy-to-follow explanation outlining the solution process

Example:

Question: A student walks from her home to the library, which is 400 meters east. Then, she walks 150 meters west to a café. What is her total displacement from her home?

Answer: 250 meters east

Explanation:

1. Define a positive direction: Choose east as the positive direction.
2. Represent movements as positive or negative values: Express the eastward walk as a positive value and the westward walk as a negative value.
3. Calculate the net displacement: Add the two values together. This will give you the overall displacement from her home.

Problem Statement:

Students often lose time manually retyping lengthy homework questions from worksheets or textbooks into search engines or AI tools.

Proposed Solution:

The AI Homework Helper eliminates this inefficiency by allowing users to upload a photo of the question directly. The application automatically extracts the content, processes it through Gemini AI, and provides a concise answer along with a detailed explanation. This approach significantly saves time and enhances student comprehension.

Inspiration:

The idea came from a simple frustration — watching students waste valuable time manually typing out long homework questions just to get help. Recognizing how much learning momentum was lost in that process, I set out to create a smarter, faster way. By combining AI with image recognition, I aimed to eliminate that barrier, helping students focus less on the busywork and more on understanding the solution. The vision was to make homework support as effortless as taking a photo — accessible, intuitive, and empowering.

What I learned:

Integrating AI APIs: Gained hands-on experience connecting to and leveraging the Gemini API to solve real-world problems.

• Image-to-Text Processing: Learned how to handle image uploads, perform content extraction (OCR), and feed results into AI models.
• User-Centric Design Thinking: Focused on reducing friction for users, emphasizing simplicity, speed, and clarity in solution design.
• Prototyping Under Time Constraints: Practiced building a functional, end-to-end prototype quickly, balancing technical feasibility with user experience.
• Problem-Solution Alignment: Sharpened the ability to identify a specific user pain point and build technology that directly addresses it.

How I built it:

I designed a lightweight app that accepts typed questions or uploaded photos, uses Optical Character Recognition (OCR) to extract text, connects to the Gemini API for problem-solving, and returns both a final answer and a step-by-step explanation — all optimized for speed, accuracy, and ease of use.

Next Step and Vision:

I believe education is more than just finding the right answer — it’s about sparking curiosity and a deeper understanding. My goal is to evolve AI Homework Helper from a simple problem-solving tool into a true learning companion that encourages students to explore why answers work, not just what they are. By delivering personalized explanations, adaptive learning paths, and interactive experiences, I aim to inspire a lifelong passion for learning.

Video link: https://vimeo.com/1077081881/7af359de1e

Built With

• c#
• google-gemini
• maui
• xaml