Java course and interview preparation, 1‑on‑1 with an engineer who ships Java every day.

The JVM runs Java bytecode. The JRE is the JVM plus the standard libraries needed to run programs. The JDK adds the tools to build them, such as the compiler and debugger. Recent Java versions no longer ship a separate JRE, which is worth mentioning.

Hash-based collections such as HashMap use hashCode() to find the bucket and equals() to find the exact key. If two equal objects have different hash codes, the map can’t find a value you put in. Equal objects must have equal hash codes.

String is immutable, so every change creates a new object. StringBuilder is mutable and faster for building strings in a loop. StringBuffer is the older synchronised version and is rarely needed today.

Checked exceptions, such as IOException, must be declared or handled at compile time. Unchecked exceptions extend RuntimeException, such as NullPointerException, and usually signal programming errors. Expect a follow-up on how you design exceptions in your own code.

Keys are hashed into an array of buckets. Collisions are stored in a list that turns into a balanced tree once a bucket gets large, since Java 8. When the map grows past its load factor, it resizes and redistributes entries. Interviewers often ask what happens with a badly written hashCode().

A synchronised map locks the whole map for every operation. ConcurrentHashMap allows concurrent reads and updates to different parts of the map without a global lock, so it scales much better under load. It also doesn’t allow null keys or values.

Spring wraps the bean in a proxy that starts the transaction and then commits or rolls it back. Because of the proxy, calling a @Transactional method from another method in the same class skips it, and by default only unchecked exceptions trigger a rollback. Both catch people out in real projects.

Loading a list of entities and then lazily loading a relationship for each one runs one query for the list plus one per item. Fix it with fetch joins, entity graphs or batch fetching, and check the SQL your ORM actually generates.

Start with metrics and logs to see where the time goes: the database, external calls, CPU, garbage collection or thread pools. Confirm with profiling and thread dumps. Interviewers want a real story: what you measured, what you found and what you changed.

Version the API, add fields rather than changing or removing them, keep error formats stable and deprecate with notice. Contract tests help catch breaking changes before release.

Usually a well-structured modular monolith first, splitting out services when there’s a clear reason, such as independent scaling or separate teams. Microservices add network failures, data consistency problems and operational cost. Senior interviews look for that judgement, not a fixed answer.

Confirm it with heap and GC metrics, take heap dumps and compare them over time to see what keeps growing, often caches, listeners or static collections. Fix the cause, then add monitoring so it’s caught earlier next time.