DEV Community

sm1ck profile picture

sm1ck

Builder of HoneyChat. Backend + infrastructure for AI companion products — persistent-memory chat, multi-tier LLM routing, GPU image/video pipelines. Background in Web3 tooling (LayerZero automation,

We Deleted 10 Real Users with a Test-Cleanup Script — RCA

We Deleted 10 Real Users with a Test-Cleanup Script — RCA

Comments
7 min read
ChromaDB 0.5 Silently Leaks Memory Until You Set One Env Var

ChromaDB 0.5 Silently Leaks Memory Until You Set One Env Var

Comments
4 min read
We Measured LLM Prompt Caching in Production — Same Prompt, 0% to 91% Hit Rates

We Measured LLM Prompt Caching in Production — Same Prompt, 0% to 91% Hit Rates

Comments
5 min read
IP-Adapter + LoRA for product catalog rendering — putting shop items on AI characters

IP-Adapter + LoRA for product catalog rendering — putting shop items on AI characters

1
Comments 1
7 min read
Character consistency in AI image generation — where prompts break down and LoRA helps

Character consistency in AI image generation — where prompts break down and LoRA helps

Comments
6 min read
LLM routing per tier via OpenRouter — when one model doesn't fit all

LLM routing per tier via OpenRouter — when one model doesn't fit all

Comments
6 min read
Building an AI companion with persistent memory — Redis + ChromaDB

Building an AI companion with persistent memory — Redis + ChromaDB

Comments
6 min read
loading...