Pinned Loading

1. lky-brain lky-brain Public

   Give a small local model (Qwen3-14B) the reasoning style of a domain expert, using a LoRA fine-tune. The expert here is Lee Kuan Yew.

   Python

2. pixi-wiki pixi-wiki Public

   Publishes Markdown KBs as a web wiki for humans and agent-readable surfaces: raw Markdown, llms.txt, and MCP.

   HTML

3. planned-program-intel planned-program-intel Public

   AI-native decision routing and institutional memory demo for enterprise event programs.

   TypeScript

4. j-space-replay j-space-replay Public

   Replay a VLM's decoded internal concept readouts synced to a video timeline (J-lens / logit-lens on Qwen2.5-VL)

   Python 1

5. effective-state-load effective-state-load Public

   Effective State Load (ESL = SC × DD) predicts when LLM agents lose world-state fidelity.

   Python

6. portfolio portfolio Public

   Personal Portfolio

   TypeScript