Turn wild ideas into things that work
Back-end by training, now building two LLM-powered products: an AI chef that decides what is for dinner, and an app that archives my hometown dialect. Here I document the whole journey of turning ideas into products — pitfalls included.
Agents that execute code need isolation, but isolation costs overhead. This breaks down six sandboxing approaches — containers, gVisor, Firecracker, ZeroBoot, Apple container, and Wasm — so you can pick the right trade-off for your threat model.
The hardest part at home isn't cooking, it's deciding what to eat. I built a small module that recommends tonight's dish from a photo of the fridge. A development note for something not yet shipped.
When an agent takes on a task like "help me make a market-research report," the hard part isn't execution but decomposing the task and managing state. This records how I used langchain4j to implement a task tree, dependency checks, retries, and reasoning traces.
A commonly conflated concept: Function Calling is a capability of the model itself, MCP is the protocol connecting model and tools. Clarifies their real differences in tool discovery, security boundary, and portability, and how Spring AI unifies the two.
Every new AI session has to re-understand your project. CLAUDE.md is the project-level memory that solves this. Of the seven sections in this template, the most useful one I've found is "Do not change."
After splitting a document, not every chunk is worth keeping. Ads, nav bars, disclaimers, and duplicates are noise that pollutes retrieval. This walks rule-based, structural, semantic, and LLM filtering strategies, with a Java pipeline you can adapt.
Xuelao is a Hoklo dialect around Shanwei, with no standard writing, no teaching resources, and young people no longer speaking it. I built a translation and learning app — harder than I expected.
Agent Skills package domain know-how and workflows into a folder an agent can load on demand. This explains how the three-layer progressive disclosure works, how code execution fits in, and how Skills differ from MCP and multi-agent architectures.
