Java Backend Developer
I build Java backend services and reusable components, focusing on cache behavior, asynchronous execution, and service boundaries. My work also spans frontend integration and AI-assisted full-stack development.
Blog · Projects · Notes · 简体中文
A Spring Cache extension for composing Redis cache protection.
- Composable policies. An ordered, configurable handler chain separates protection rules from cache operations.
- Concurrent loading. With
sync=true, same-key callers within one JVM share a loading result; cross-instance coordination requires a distributed-lock backend.
Stack: Java · Spring Cache · Redis · Redisson
Early-stage; APIs may change. The main branch targets Spring Boot 4 / Java 21; check compatibility before choosing a published release.
Loading implementation · Compatibility & limitations
An online judge platform for submitting code and evaluating it against test cases.
- Task delivery. The outbox path feeds Redis Streams, with retries for dispatch failures and dead-letter handling for malformed jobs.
- Execution boundary. Independent judge workers run submitted code in Docker with network isolation and CPU, memory, and process limits.
Stack: Java · Spring Boot · Redis · Docker · Vue 3 · TypeScript
The Core profile is an opt-in architecture experiment, not the default topology. The project has no production environment.
Delivery implementation · Project status & boundaries (Chinese)
I start from explicit contracts and failure cases, and check changes with focused tests, logs, and reproducible steps. I write and debug code independently, using AI to assist exploration and implementation while reviewing its output myself.
Selected implementation notes. Both articles are in Chinese; original titles are retained.
- 缓存击穿时，为什么要同时有 Future、分布式锁和 double-check？ — Request coalescing, lock scope, and failure propagation.
- UltiCode Outbox 与 Redis Streams：把判题投递做成可恢复状态 — Durable delivery intent, retries, and recovery boundaries.