Candidates lose interviews on questions they answered correctly. The textbook reply is not wrong — it just does not contain the thing the interviewer was listening for.
Here are six you will be asked, and what each is actually measuring.
"Explain MVC"
Surface: model, view, controller, separation of concerns.
Real: have you ever had to find a bug in somebody else's code? The reason to separate concerns is that when a page shows the wrong total, you know which of three places to open. Say that, and you have answered as someone who has maintained software rather than read about it.
"What is a database index?"
Surface: it makes queries faster.
Real: do you understand the cost? An index makes reads faster and writes slower, and takes disk. The follow-up is often "so why not index every column", and the candidate who already said "because inserts pay for it" has answered that before it is asked.

"Why shouldn't you store passwords as MD5?"
Surface: MD5 is broken.
Real: do you know what hashing is for? The useful answer is that MD5 is fast, and fast is the wrong property — an attacker with the hash file tries billions of guesses per second. You want something deliberately slow and salted, which in PHP means password_hash() and nothing you wrote yourself.
"How would you find a slow page?"
Surface: optimise the code.
Real: do you measure or guess? Say you would look at the query count first, because in most slow pages the answer is a loop issuing one query per row rather than one query for all of them. Naming the N+1 problem by behaviour rather than by name shows you have met it.
"Tell me about a bug you could not fix"
Surface: a story where you fixed it anyway.
Real: will you tell your lead when you are stuck? The worst answer is that you have never been stuck. The best describes how long you worked alone before asking, and what you had already ruled out when you did — because the cost of a junior is rarely the mistakes, it is the silence.
"Do you have any questions for us?"
Surface: no, you covered everything.
Real: this is still the interview. Ask what the first three months look like, who reviews your code, and what went wrong on the last project. The third question is the one that gets an honest answer, and it tells you whether you want the job.
The pattern underneath all six
Every one of these rewards the same thing: evidence that you have done the work rather than read about it. That is not a trick you can learn the night before.
It is, however, exactly what the project in part one gives you. If you built something, deployed it and watched somebody use it badly, you will have a real answer to all six without preparing any of them.
Part three, tomorrow: putting a Laravel application on cPanel shared hosting, which is where most of these projects die.




