Curios Campus: AI-Powered Self-Tutoring Platform

Project Name
Curios Campus: EdTech Innovation Platform

Project Type
AI-Integrated Persona-Based Learning Platform

Project Scope

Open-Source AI Models, Custom Embedding Logic, Scalable Cloud Architecture

Duration
1+ Year with Ongoing Status

Objective
The idea of creating is to have a platform where the teachers upload the syllabus material and the exams are generated by the system for students. These exams keep changing with the student performance level, increased or decreased pace of learning, and time. Hence, the system should generate multiple-choice question-type examinations and math-based image problems for both educational and technical learning environments. The idea of this product is to work as a highly personalized learning companion that performs assessments in real time and gives comments.

Client Background & Initial Challenges

The client was an EdTech startup with the philosophy that AI could have an enterprising role in personalized education. Now, while the concept itself was strong, technically it had to be utterly razed and built from scratch. Some initial hurdles were confronted in this task:

Embedding engine missing
Created the personalized learning algorithm from scratch.
Model integration
Getting DeepSeek and OpenAI models implemented to generate dynamic context-aware questions from syllabus input was a hard task.
Migration Issues
The backend and database were so ancient that there had to be a migration to a much more solid and scalable platform.

Goals & KPIs

The project goal was to conceive a smart and scalable learning ecosystem with seamless interactions supported by tutor-like AI features.

Goals

Develop modern, intuitive UIs that work well for students and teachers alike.
Build and customize an AI embedding engine for real-time student assessment and for determining the difficulty level of questions appropriately.
Generate exams on the basis of defined syllabuses with real-time modifications.

Key Performance Indicators (KPIs)

Ques: How closely related are the questions to the syllabus?
Answer Match: More concerned with complicated math questions.
Image Quality: Generated images for math questions should be accurate, clear, and relevant.
Student Dashboard Metrics: Real-time reports about the progress and performance of a student.

Technical Architecture & Tools

To ensure the platform is scalable, modular, and cloud-native, a modern tech stack was deployed

Core Technology Stack

Solution Components

A modular, scalable solution was designed to address all aspects of the platform:

Progressive Web Application (PWA):

A high-speed, mobile-responsive interface for both students and teachers with an app-like feel.

AI & Embedding Engine

Developed a sophisticated AI module for question generation, student evaluation, and adaptive learning based on performance. Integrated DeepSeek, OpenAI, and custom logic for embedding and real-time decision-making.

Microservices Architecture

Enabled core functions of the platform (exam generation, evaluation, analytics, and image rendering) to run independently and efficiently across services.

Development Process

The platform was built using a robust Agile methodology and a DevOps-driven CI/CD pipeline, ensuring fast iterations and stable releases.

Weekly sprint planning and demo reviews with the client team
Continuous integration using automated testing pipelines
Smooth deployment and rollback using DevOps best practices

Impact & Ongoing Development

Though still evolving, the platform has demonstrated early success in

Personalizing learning at scale, with real-time adaptive exams

Reducing manual workload for teachers by automating exam generation

Increasing student engagement by delivering dynamic, challenging, and interactive learning experiences

Providing deep learning analytics to identify student strengths and areas for improvement

Curious Campus is a next-generation AI-powered self-tutoring solution that transforms how students learn and how teachers evaluate. Built with a strong foundation in embedding technology, adaptive learning logic, and a cloud-native framework, this platform is set to redefine EdTech for institutions seeking personalization, automation, and insight-driven instruction.