I'm currently learning about AI and exploring how it fits into useful software. My interests span frontend development, backend development, and the ideas behind AI agents.
My current focus is dsh-lark-agent and Latch.
|
Reach your DeepSeek Harness assistant from Feishu / Lark. Continue conversations, hand off tasks, and respond to confirmations from your phone. |
|
Coming soon. |
|
Evidence-backed review for Agent Skills, with paired evaluations and inspectable execution traces. |
|
Parse, check, and format TOML with Rust — through a library, a CLI, and an LSP server. |
More experiments
- dsh-session-tree — exploring session lineage, forks, and subagents in DeepSeek Harness.
- lottie-miniprogram-v2 — a Canvas v2 Lottie player for Douyin Mini Apps and compatible runtimes.
| Area | Technologies & interests |
|---|---|
| Frontend | TypeScript · JavaScript · HTML · CSS · React · Vue |
| Backend | Node.js · Rust · APIs and services |
| AI (exploring) | LLM applications · Agent workflows · Tool use · Evals |
- AI & LLMs — the fundamentals and how to build practical applications around them.
- Agents — how they use tools, manage context, and work through tasks.
- Putting it together — connecting interfaces and backend services with AI capabilities.
I learn through small experiments, hands-on practice, and following my curiosity.
Keep learning. Keep building.