# Project Overview
Developed and maintained a English-learning platform offering voice and video lessons for Korean learners.
The system supported lesson scheduling, payment, teacher management, and AI-assisted feedback for students.
----
# Team Information
Full-stack team of 2 in-house developers.
Collaborated closely with the design teams to align development with product goals.
----
# Development & Implementation A — API Modernization
## Overview:
Refactored a monolithic Classic ASP system into a modular architecture to improve maintainability and scalability.
## Feature Developed:
Rebuilt APIs for scheduling, report generation, and teacher/student management.
## Challenges:
Code was not reusable, resulting in excessive duplication and inconsistent data handling across modules.
Feature improvements required modifying multiple redundant code paths, slowing down release cycles.
## Approach & Technologies:
Initially refactored Classic ASP logic into entity-based classes with internal DB access functions.
Later, separated APIs into ASP.NET MVC, implementing clear Controller–Model–View layers for better maintainability.
## Results:
Reduced code duplication and improved feature development speed, enabling faster iteration and integration with new systems.
----
# Development & Implementation B — WebRTC-based Lesson Integration
## Overview:
Implemented WebRTC-based video lesson functionality to replace the costly VoIP system and support growing demand for online video lessons.
## Feature Developed:
Developed token-based authentication, and real-time connection APIs.
## Challenges:
The existing VoIP solution had high licensing costs and lacked flexibility for video-based learning expansion.
## Approach & Technologies:
Translated and shared English technical documentation with the team, conducted internal reviews,
and created a working prototype for proof of concept.
This project marked the start of adopting ASP.NET MVC for API separation and migration.
## Results:
Reduced per-session costs by 80% and improved lesson connection success rates.
----
# Development & Implementation C — AI Feedback Automation
## Overview:
Developed AI-assisted writing correction and post-class feedback systems using text and voice data.
## Feature Developed:
Implemented AI-based writing correction for student essays and feedback generation from VTT subtitles of recorded lessons.
## Challenges:
Aimed to reduce teachers’ post-class workload while enriching the student’s learning outcome with automatically generated materials.
## Approach & Technologies:
Integrated Gemini API with Google Cloud TTS and STT services to process lesson transcripts and generate contextual feedback.
## Results:
Reduced manual workload for teachers and enhanced the richness of post-lesson materials, improving overall user satisfaction.