ChatnLearn EdTech

Inspiration

Inspired by Cambridge's research-based pedagogy — Oracy, it is believed that the key future skill, as highlighted during the Hackathon Alpha workshop, is the ability to communicate effectively.

A future is envisioned where AI does not replace humans, but rather empowers performance and learning. For this reason, ChatnLearn was built. We aim to address educational challenges such as large class sizes, which limit individual student participation, and the unaffordability of private tutoring for underprivileged learners.

What It Does

ChatnLearn offers two interactive learning modes:

Talk to Learn: Subject knowledge (e.g., physics) is integrated into speaking practice.
Learn to Talk: Fluency and confidence in language use are developed.

A web-based platform is provided where students can select a topic and join guided speaking sessions with an AI avatar.

Prior to the session, students review preparation materials that include:

Simplified explanations of complex concepts
Key vocabulary
Useful sentence structures

Practice is then conducted with the AI tutor, followed by a personalized performance report offering constructive feedback.

How We Built It

The capabilities of Large Language Models (LLMs) are leveraged to enable personalized, oracy-based learning. The conversational AI architecture integrates:

Speech-to-Text
LLM processing
Text-to-Speech
Intelligent conversation logic

Various APIs are currently used to implement these features. The next step involves evaluating solutions for economic sustainability and scalability.

Challenges We Ran Into

Despite the apparent readiness of speech-to-speech technology, multiple challenges were encountered:

Significant time was spent mitigating latency issues
Synchronization for lip movement proved complex
Exporting real-time transcripts was particularly difficult

These obstacles required extensive testing and optimization.

Accomplishments We’re Proud Of

Initial product testing was conducted with users, and it was noted that the solution significantly encouraged speaking practice. Positive anticipation was expressed for a more comprehensive learning journey, including expanded materials and AI-generated feedback.

What We Learned

Key lessons were gained on how conversational AI integrates multiple technologies. Additionally, hands-on experience was obtained in collaborative and AI-assisted development, using tools such as:

ChatGPT
Duobao
Cursor

The capabilities of these tools were found to be highly impressive.

What’s Next for ChatnLearn EdTech

Though initially targeted at secondary school students in Hong Kong, ChatnLearn is seen as having global potential across age groups and regions.

Unexpected interest was shown by parents, indicating broader appeal. Curiosity about AI — a buzzword often not well understood — led to valuable feedback and new directions.

Therefore, exploration will be made into how the product can be expanded to support: