Skip to content

FEAT: Subject-aware learning system with mode-based reasoning + docs#10

Merged
lingarobotics merged 3 commits intomainfrom
feat/context-aware-prompts
Mar 28, 2026
Merged

FEAT: Subject-aware learning system with mode-based reasoning + docs#10
lingarobotics merged 3 commits intomainfrom
feat/context-aware-prompts

Conversation

@lingarobotics
Copy link
Copy Markdown
Owner

🚀 Overview

This PR introduces a subject-aware learning system integrated across all modes, along with a structured documentation layer and UX improvements.

The system shifts from generic AI responses to context-controlled, learning-oriented outputs.


🎯 Why

Previously, responses were:

  • Generalized across domains
  • Not aligned with exam context
  • Lacking structured learning flow

This PR solves that by introducing:

  • Subject-aware reasoning
  • Mode-based behavior control
  • Guided learning flow for users

🧠 Core Features Added

🎯 Subject-Aware Learning

  • Users can define a subject (e.g., Automotive Mechatronics)
  • Responses are strictly constrained within that domain
  • Eliminates cross-domain leakage

📌 Core Points Extraction

  • Converts full explanations into exam-ready key points
  • Helps in revision and answer structuring

🧠 Understanding Verification

  • Implemented via Teach-Back Mode
  • Users explain → system evaluates correctness and gaps

🧩 Structured Answer System

  • Enforces:
    • Scope awareness
    • Aspect identification
    • Logical flow
  • Produces exam-oriented responses

🧠 Mode Behavior (Refined)

  • Learn Mode → Full structured, exam-ready explanations
  • Doubt Mode → Precise, scope-limited clarification
  • Fast Learn Mode → Key points for quick recall
  • Teach-Back Mode → Understanding verification

📘 Documentation System

  • Added FEATURE.md for core system explanation
  • Implemented /docs page:
    • Markdown-style structured UI
    • Sticky navigation (desktop)
    • Mobile-friendly layout
    • Mode selection guidance + learning flow

🎨 UX Improvements

  • Added Docs navigation in header
  • Introduced Subject Input component across modes
  • Improved layout consistency
  • Fixed reset-password API payload mismatch

🏗️ Implementation Details

Backend

  • Updated prompt system for strict subject enforcement
  • Extended API to accept context
  • Strengthened response control logic

Frontend

  • Added ContextProvider for global subject state
  • Created reusable SubjectInput component
  • Integrated context across all modes
  • Built Docs page with structured layout

🧪 Testing Notes

  • Verified subject restriction across modes
  • Checked fallback behavior (no subject → general response)
  • Validated UI responsiveness (desktop + mobile)
  • Tested reset-password flow

🧭 Impact

This PR transforms the system from:

“AI that answers questions”

to:

“AI that guides structured learning and verifies understanding”


🏁 Summary

  • Introduces subject-aware reasoning
  • Strengthens mode-based learning system
  • Adds structured documentation
  • Improves UX and consistency

Why:
Users need clarity on how to use different modes and understand system capabilities

What:
- Added FEATURE.md documenting core features and learning system design
- Implemented Docs page with markdown-style structured layout
- Introduced mode selection guidance and recommended learning flow

How:
- Built Docs.jsx with sectioned layout and sticky navigation
- Designed docs as a decision guide instead of static feature listing
Why:
General AI responses were leaking across domains, reducing clarity and exam relevance

What:
- Introduced subject input for focused responses
- Integrated context across all modes
- Enforced strict domain-aware prompting

How:
- Added ContextProvider for global state
- Created SubjectInput component
- Passed context through API pipeline
- Strengthened prompt rules for subject alignment
Why:
Improve usability and fix API mismatch

What:
- Added Docs navigation
- Fixed reset-password payload
- UI improvements

How:
- Updated AppShell
- Aligned frontend-backend contract
- Refined styles
@vercel
Copy link
Copy Markdown

vercel Bot commented Mar 28, 2026

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
lgc-concept-ai Ready Ready Preview, Comment Mar 28, 2026 8:50pm

@lingarobotics lingarobotics merged commit b2e42af into main Mar 28, 2026
3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant