EmotiLearn AI

• 

  Logo

Inspiration

In today’s fast-paced digital world, emotional intelligence (EQ) is more essential than ever yet it’s often overlooked in traditional education. As an educator and AI researcher, I wanted to create something beyond rote learning: a tool that fosters empathy, reflection, and emotional decision-making in a gamified and interactive way. That vision led to the birth of EmotiLearn AI a real-time, generative AI-powered game that helps teenagers grow their emotional skills through dynamically generated scenarios.

What it does

EmotiLearn AI is an interactive emotional intelligence game that: Uses Groq's Llama 3 model to generate unique real-life scenarios. Asks users to choose the best emotional response from multiple options (A–D). Provides immediate, motivational feedback. Tracks and displays the emotional intelligence score to encourage self-awareness and growth. Offers a fresh experience each time, no two questions are the same.

How we built it

Frontend/UI: Streamlit framework was used for its lightweight, responsive, and interactive interface. Backend Intelligence: Integrated Groq API with Llama 3-70B model to generate contextually rich and emotionally nuanced scenarios. Randomized correct options each time to avoid predictability. Logic: Each question is structured with a scenario, four choices, a correct answer, and an explanation. Users are scored and motivated based on their responses. Hosting: Deployed on Hugging Face Spaces using a Streamlit app. Security: Managed API keys via secure environment variables.

Challenges we ran into

Model Limitations: Groq’s older mixtral model was deprecated during development, requiring quick migration to llama3-70b-8192. Hugging Face Static Space Limitations: Static Spaces didn’t support secrets, so we had to use a Streamlit app Space instead. Prompt Engineering: Crafting effective prompts to ensure the model generated scenarios with clear multiple-choice answers. Randomization Bugs: Initially, the same correct option (e.g., always "C") was selected. We fixed this by implementing proper shuffling logic. Balancing Feedback: Making sure feedback was constructive, not judgmental, especially for a younger audience.

Accomplishments that we're proud of

🎮 Created a fully functional emotional intelligence game using live AI generation. 🧠 Enabled replayability - every scenario is unique and fresh. 🚀 Successfully deployed on Hugging Face Spaces with working Groq integration. 🧩 Built a gamified self-improvement tool aligned with real-life SEL (social-emotional learning) goals. 🧑‍🏫 Used cutting-edge AI for meaningful education, not just entertainment.

What we learned

How to integrate LLMs like Groq’s Llama 3 with Streamlit apps. Effective prompt design to structure complex, empathetic questions from language models. UX for learning apps: Keeping users engaged through motivational design. Handling environment variables and deployment constraints in Hugging Face Spaces. Why EQ matters: Even AI can help nurture it when designed thoughtfully.

What's next for EmotiLearn AI

🔊 Add voice narration and background music to increase immersion. 🌍 Translate to local languages (e.g., Urdu, Spanish) for global impact. 📱 Export as Android/iOS app via WebView or Flutter wrapper. 👥 Add multiplayer mode for classroom EQ competitions. 📊 Build a dashboard for educators to monitor emotional development progress. 🧪 Conduct real-world trials in schools and with counselors to improve and validate the game's impact.

Built With

• groq-api-(llama-3)
• hugging-face-spaces
• markdown
• streamlit