talantir

Leaderbaord
Primary student Dashboard
Analytics
Simulations ,
Demo of IIP

VIDEO IS ON GITHUB

Inspiration

Modern classrooms are still largely built around one-way teaching, despite students learning in completely different ways. Some students are visual learners, some understand through interaction, while others need personalized explanations or more engagement to stay focused.

We noticed that while schools have smart boards and digital tools, they are usually disconnected, static, and not context-aware. Teachers often switch between presentations, videos, attendance systems, quizzes, and assignments manually, which interrupts the flow of teaching. At the same time, students rely on fragmented platforms like Teams or WhatsApp that were never truly designed for physical classrooms.

We wanted to build an AI-native classroom system that acts like a real-time educational companion , one that listens to the lecture, understands the topic being taught, and instantly transforms it into interactive visual learning experiences while also supporting teachers, students, and school administrators through a unified platform.

What it does

Talantir is an AI-powered classroom and adaptive learning platform built as a supplement to traditional education.

The platform has two major components:

1. AI Smart Classroom System (Interactive Panel Layer)

Talantir runs directly on classroom interactive panels (IIPs) and listens to live classroom discussions in real time. Using speech recognition and contextual AI understanding, it dynamically displays the most useful educational content for the current topic being taught.

Depending on the subject, Talantir can:

Launch interactive simulations from platforms like PhET
Display diagrams, GIFs, images, and short videos
Generate real-world visualizations of concepts
Answer student questions contextually
Create live summaries and revision notes
Generate presentations for future reference
Record homework, assignments, and class to-dos

The classroom assistant also supports:

AI-generated quizzes and polls
Automatic attendance
Classroom monitoring and engagement tracking
Participation analytics
Focus and attention scoring
ADHD-friendly calming visuals and adaptive patterns
Voice-based interaction with students and teachers

2. Unified School Dashboard

Talantir also includes a complete multi-role dashboard system for:

Students
Teachers
Administrators

The dashboard acts as a centralized school platform replacing fragmented systems currently spread across apps like Teams, WhatsApp, and parent portals.

Features include:

Assignments and submissions
Practice tests
AI-generated class summaries
Reminders and notifications
Student feedback systems
Performance analytics
Participation tracking
Learning insights for teachers

Teachers receive AI-generated recommendations such as:

Which students need additional support
Which students are highly engaged
Participation improvement suggestions
Classroom trend analysis

Students get access to an advanced AI tutor:

A high-context, multimodal AI tutor capable of voice interaction, whiteboard explanations, simulations, videos, and adaptive teaching styles for competitive exam preparation and personalized learning.

How we built it

We designed Talantir as a modular AI-first platform combining:

Real-time speech recognition
Context-aware LLM pipelines
Retrieval-based educational grounding
Interactive classroom UI systems
Dashboard analytics
Adaptive learning workflows

The architecture combines:

Live audio transcription
Topic extraction pipelines
Educational content retrieval
Simulation orchestration
AI tutoring agents
Classroom analytics engines

For the frontend, we focused heavily on creating a clean, distraction-free interface optimized for real classroom usage.

The dashboard ecosystem was designed around role-based workflows so students, teachers, and administrators each get interfaces tailored to their needs.

The AI tutor system was built around multimodal interaction:

Voice
Whiteboard explanations
Simulations
Visual aids
Adaptive conversational tutoring

Challenges we ran into

One of the biggest challenges was designing a system capable of understanding live classroom context in real time without overwhelming either the teacher or students.

Another major challenge was balancing automation with usability. Many educational tools become too complicated for teachers to use naturally in classrooms. We had to ensure Talantir enhanced teaching instead of interrupting it.

We also faced challenges with:

Real time latency
Educational content grounding
Topic relevance filtering
Attention analysis
Multimodal synchronization between audio, visuals, and simulations

Designing analytics that were actually meaningful, instead of just generating random metrics was another important challenge.

Accomplishments that we're proud of

We are proud that Talantir reimagines the classroom as an intelligent, adaptive environment instead of just digitizing traditional teaching methods.

Some accomplishments we are especially proud of:

In our School , we approached the principal and pitched the idea to her, she loved the idea and decided to demo it with the classrooms the coming year. were also proud of the building process where we:
Built a real-time AI classroom assistant concept
Combined simulations, tutoring, analytics, and dashboards into one ecosystem
Designed accessibility-focused learning support
Created a platform that supports multiple learning styles
Built an AI tutor capable of multimodal interaction
Replaced fragmented school tools with a unified platform vision

We are also proud of how scalable the idea is from schools and coaching centers to hybrid learning environments.

What we learned

Through building Talantir, we learned that educational technology should not replace teachers it should amplify them.

We also learned how important contextual grounding is for AI in education. Generic AI responses are not enough in classrooms; the system must deeply understand what is currently being taught and adapt instantly.

On the technical side, we learned:

Real-time AI orchestration
Educational retrieval systems
Designing low-distraction interfaces
Multimodal AI interaction
Analytics pipeline design
Human centered classroom UX

Most importantly, we learned that engagement and understanding matter far more than simply delivering information.

What's next for Talantir

Our next goal is to turn Talantir into a fully deployable classroom intelligence platform.

Future plans include:

After the demo we plan to apply it to all the classrooms in school and then surrounding schools too.
Deeper integration with school hardware
Add more languages to support
Emotion and engagement-aware teaching adaptation
Personalized learning pathways
AR/VR-powered educational simulations

We also want to build a scalable ecosystem where schools can plug in custom educational content, simulations, and teaching modules tailored to their curriculum.