AI Briefing
Karrot news
The latest AI and developer news from Karrot, with short summaries.
Visit website →What We Learned Using 2 Trillion Tokens for Category Classification
Danggeun built an LLM-based Taxonomy Management System to automatically classify categories and attributes at scale.
How Should Long-Term User Modeling Change in a Local Super App
User embeddings trained on long-term logs, combined with RCBS, significantly boosted recommendation performance.
Solving a 200MB Module at the Team Level: Karrot's Short-form Team Adopts On-demand Dynamic Feature Module
Karrot's short-form team split off **40MB** of editing features into an **On-demand DFM**, reducing the burden on global users.
Improving iOS Webview Input Experience for Webview Engineers
We worked around the keyboard-push issue in iOS Webview using an opacity trick.
Redux on the Server: Building a Node.js Event Sourcing Library
To overcome the limits of CRUD, event sourcing and Ventyd for TypeScript were introduced.
The Evolution Journey of Danggeun Pay's Backend Architecture
A journey that evolved from Layered to Hexagonal, Clean Architecture, and Monorepo.
Building Event Center: Karrot's User Event Management Platform
Karrot shifted from code-centric event management to a UI-based Event Center with a type-safe CLI.
Karrot's User Behavior Log Management Platform: Building the Event Center
By turning code-managed user behavior logs into a UI-based platform, Karrot improved consistency and automation.
Standardizing User Activation: How Karrot Built a Shared Data Layer
Karrot built an organization-wide shared data layer to ensure consistency and reliability in user activation metrics.
3 Problems Solved by Building Activation as a Company-Wide Common Layer: Reliability, Cost, and Productivity
Daangn built User Activation as a company-wide common data layer to improve data reliability and operational efficiency.