Inspiration ## Inspiration
Traditional education follows a one-size-fits-all approach, leaving many learners behind. People learn differently—some are visual, others are auditory or kinesthetic. We were inspired to build an AI agent that doesn't just deliver content, but actively discovers how a user learns best and dynamically adapts to their unique style.
What it does
CogniLearn AI is an adaptive, diagnostic AI tutor. Instead of asking users to self-identify their learning style, the agent presents material through multi-modal formats (visual diagrams, audio summaries, and interactive micro-simulations). It analyzes user interaction, time spent, and response accuracy to build a dynamic cognitive profile. The agent then personalizes future lessons, quizzes, and continuous learning journeys around the user's optimal style.
How we built it
- AI Engine: Built with Large Language Models (LLMs) integrated via Nebius infrastructure.
- Frontend: Streamlit / React interactive interface for multi-modal rendering (diagrams via Mermaid.js, audio generation, and quiz modules).
- Agentic Logic: Custom Python workflow tracking engagement metrics to auto-calibrate learning personas.
Challenges we ran into
- Balancing implicit user diagnostic tracking with a smooth, non-intrusive UI/UX.
- Dynamically generating structured visual flowcharts and interactive scenarios in real time.
Accomplishments that we're proud of
- Successfully creating an agent that shifts from passive content generation to active diagnostic personalization.
- Enabling real-time adaptation across visual, auditory, and practical learning formats.
What we learned
- How to combine multi-modal AI feedback loops into a unified personalized agent experience.
- Designing diagnostic UI patterns that gather learning behavioral data naturally without user fatigue.
What it does
How we built it
Challenges we ran into
Accomplishments that we're proud of
What we learned
What's next for CogniLearn Ai
Built With
- ai-agents
- llm
- nebius
- python
- streamlit
Log in or sign up for Devpost to join the conversation.