LTD AI Co-Teacher
Turning a Learning Framework into an AI Co-Teacher
INDUSTRY
SeqHub product
capability
Custom Products
Look-Think-Do is a learning framework built on the idea that students learn to code by analysing existing code, predicting what changes will do, and then building independently. As a teaching method it works. As software it did not exist.
The Challenge
Most AI coding tools answer questions by writing the code. That is exactly what a learning platform must not do. The system needed to guide without solving, assess understanding at each stage, and give three different kinds of user, learners, instructors, and parents, a view suited to them.
It also needed to run real code in the browser, respond in real time, and stay usable for students at very different levels.
What We Built
Three-phase framework. Look: interactive code walkthroughs with comprehension checks. Think: students hypothesise changes and predict outcomes before testing them. Do: independent building with direction-only prompting.
In-browser code execution, with a three-panel interface for guidance, code, and output.
Highlight-and-ask. Students can select any section of code and ask a question about it directly.
Real-time streaming responses with context awareness across all three phases.
Role-based access for learners, instructors, and parents, with progress visibility appropriate to each.
Per-phase assessment, tracking comprehension accuracy, prediction quality, and code quality separately.
Project templates mapped to curriculum chapters, held in a cloud repository.
Session continuity across browser tabs, with undo and iteration history.
The Outcome
In active development and real-world use within SeqHub's learning programmes.
The Challenge
Most AI coding tools answer questions by writing the code. That is exactly what a learning platform must not do. The system needed to guide without solving, assess understanding at each stage, and give three different kinds of user, learners, instructors, and parents, a view suited to them.
It also needed to run real code in the browser, respond in real time, and stay usable for students at very different levels.
What We Built
Three-phase framework. Look: interactive code walkthroughs with comprehension checks. Think: students hypothesise changes and predict outcomes before testing them. Do: independent building with direction-only prompting.
In-browser code execution, with a three-panel interface for guidance, code, and output.
Highlight-and-ask. Students can select any section of code and ask a question about it directly.
Real-time streaming responses with context awareness across all three phases.
Role-based access for learners, instructors, and parents, with progress visibility appropriate to each.
Per-phase assessment, tracking comprehension accuracy, prediction quality, and code quality separately.
Project templates mapped to curriculum chapters, held in a cloud repository.
Session continuity across browser tabs, with undo and iteration history.
The Outcome
In active development and real-world use within SeqHub's learning programmes.