Baemin Frontend Developer's Adaptation to Platform Organization: Data-Driven Prioritization and AI Automation

A frontend developer from Baedal Minjok shares their experience of moving from a service organization to a platform organization, adapting to the unfamiliar environment, and redesigning their workflow. Initially, they felt overwhelmed managing Exhibition Admin, which handled over 50 menus, while working alongside server developers without designers or PMs.

Data-Driven Problem Definition and Prioritization

To collect user behavior data, they designed a standardized Tracker module and built a dashboard. Log analysis revealed that filter manipulation sessions accounted for over 80%, leading to the introduction of recent condition restoration and filter tab features to improve this. As a result, the rate of direct navigation to detail pages increased from 5.8% to 22.6%, approximately 3.9 times, and the median access time decreased from 10.4 seconds to 8.3 seconds, a reduction of over 20%. Additionally, data-driven prioritization achieved 82% coverage of all admin features.

Workflow Efficiency Through AI and Automation

Facing a shortage of dedicated staff, the developer had to handle various roles including QA, data analysis, and policy review, and used AI tools to reduce their workload. They connected BigQuery via MCP and utilized LLM Wiki and Skills to automate data schema exploration and policy reference. In the pre-deployment verification stage, they configured AI to execute test cases (TC) in real browsers using Playwright, automatically running 96 TCs in a single deployment to ensure stability. Building such systems reduced repetitive tasks and established a sustainable work environment.