Explain concurrency and reliability tradeoffs

The non-design fundamentals rounds covered several topics: 1. **Thread-safe counter** - Implement an `AtomicCounter` or similar class with `increment()` and `get()` using built-in synchronization primitives such as `synchronized`. - Explain why the implementation is correct and what its contention limits are. 2. **Double-checked locking** - Explain how lazy initialization with double-checked locking works. - Describe why a `volatile` reference is needed, what can go wrong without it, and when this pattern is appropriate. 3. **Distributed systems operations** - Given a large-scale data platform service, explain how you would identify bottlenecks and balance consistency, availability, and scalability. - Discuss when concepts such as CDN, consistent hashing, sharding, backpressure, and failover are useful. - Walk through likely failure scenarios for systems similar to Kafka pipelines, Spark jobs, and job schedulers.