MiniWallet project
One e-wallet that grows with the roadmap. Each core post has a 30–45 minute hands-on lab; you write the core, an agent only handles scaffolding.
1
Write it
Type the core yourself before asking an agent.
2
Run it
See the output the lab shows.
3
Break it
Make it wrong on purpose and watch it fail.
4
Explain it
Two minutes out loud, without the code.
5
Get grilled
Paste the three questions to an agent.
Phase 0
Plain Java, no framework
Money, accounts and an in-memory transfer
Java 21 · JUnit 5
- Object-oriented programming: immutable objects, equals/hashCode and final0/6 steps
- Object-oriented programming: inheritance, encapsulation and polymorphism0/6 steps
- Java exceptions: checked vs unchecked, try-catch-finally and the call stack0/5 steps
- Java generics: wildcards and the PECS rule0/5 steps
- Inside Java Collections: ArrayList, LinkedList, ArrayDeque and HashMap0/5 steps
- Unit testing with JUnit 5 and Mockito: test decisions, not wiring0/5 steps
- Java concurrency: race conditions, volatile and deadlocks0/5 steps
- ExecutorService, thread pools and virtual threads in Java0/4 steps
Phase 1
One Spring Boot service
Open account, transfer and history APIs with correct transactions and locking
Spring MVC · JPA · PostgreSQL · Flyway
- REST APIs: REST constraints, API design conventions and Spring Boot project structure0/5 steps
- IoC and Dependency Injection: how the Spring container works0/5 steps
- Databases: transactions, isolation levels and locking0/6 steps
- ORM and JPA: JPQL, native queries, @Transactional and locking0/5 steps
- Optimistic or pessimistic locking: measured on a 100-seat booking0/5 steps
- Connection pooling and HikariCP: internals, settings and sizing0/5 steps
- ORM and JPA: the persistence context, dirty checking and N+10/5 steps
- Databases: indexes, EXPLAIN, N+1, pagination and soft delete0/5 steps
- How SQL runs and how to read a PostgreSQL execution plan0/4 steps
Phase 2
Make it reliable
Retries never move money twice; the cache never serves a stale balance
Idempotency-Key · Outbox · Redis · Rate limit
- Idempotency part 1: at-least-once, unique constraints and an Idempotency-Key on PostgreSQL0/5 steps
- Distributed transactions: 2PC, TCC, saga, pivot transactions and the transactional outbox0/6 steps
- Cache-aside, write-around and keeping the cache consistent with the database0/5 steps
- Cache failures: avalanche, breakdown, penetration, hot keys and big keys0/5 steps
- Rate limiting: five algorithms, a token bucket on Redis, and designing a distributed limiter0/5 steps
Phase 3
Split into services
Top-up through a fake gateway, late or duplicate IPN, saga
payment-service · notification-service · Kafka
Phase 4
Gateway and realtime
JWT and rate limiting at the edge; push balance updates to the browser
Spring Cloud Gateway · WebFlux · SSE
Phase 5
Operate
Run the whole system, observe it and load-test it
Docker Compose · CI · Tracing · k6
- The system design interview: a four-step process and clarifying requirements0/5 steps
- Back-of-the-envelope estimation: numbers to know and how to calculate fast0/5 steps
- Design principles: CAP, PACELC, scalability, performance and availability0/5 steps
- Analysing trade-offs like a senior: ideality, sweet spots, coupling and separation0/5 steps
- Sharding in system design: choosing a shard key, four case studies and resharding0/5 steps