Engineering Blog
Principles and lessons from building and running AI, RAG and media-processing systems in production. No made-up case studies — only what we've actually run.
Caching vs. Semantic Caching in AI Chatbots: What's the Difference?
If an AI chatbot answers the same question from scratch every time, cost and response time pile up. Here is how a regular cache (exact match) and a semantic cache (similar meaning) differ, their pros and cons, and what to watch out for when running them.
- #AI chatbot
- #Caching
- #Semantic cache
- #Embeddings
- #LLM cost
Notes on Transcoding from Building a Content Management System
Transcoding is where a content management system (CMS) for video, photos and audio spends most of its time and resources. Here are the principles we settled on while building one — keeping originals untouched, purpose-built copies, quality-based encoding, watermarking, camera RAW and operating with failure in mind.
- #Transcoding
- #CMS
- #Video processing
- #HLS
- #Watermarking
- #FFmpeg
Hybrid RAG: Why Vector Search Alone Isn't Enough
Hybrid RAG combines semantic (vector) search and keyword search so each covers the other's blind spots. Here we cover the concept, its benefits, what to watch for in Korean-language services, and how to decide whether you need it.
- #RAG
- #Hybrid search
- #Vector search
- #Keyword search
- #Korean search