Mi ne scias, kiom malfrue li alvenos.

Because scii means to know a fact, piece of information, or answer.
Koni means to be acquainted with or to know a person, place, or thing through experience.

Here, the speaker does not know some information: how late he will arrive. So scii is the correct verb.

• Mi scias la respondon = I know the answer.
• Mi konas lin = I know him.

Here kiom means how much / to what degree. In this sentence, it is being used with malfrue to mean how late.

So kiom malfrue literally means something like to what degree late.

This is a very normal Esperanto pattern:

• kiom rapide = how quickly
• kiom longe = how long
• kiom forte = how strongly

Because the word is describing how he will arrive, so it needs to be an adverb, not an adjective.

• -a = adjective, used for nouns
• -e = adverb, used for verbs, adjectives, or other adverbs

Here malfrue modifies alvenos:

• li alvenos malfrue = he will arrive late

If you used malfrua, you would need a noun:

• malfrua trajno = a late train

It breaks down like this:

• fru- = early
• mal- = opposite
• -e = adverb ending

So:

• frue = early
• malfrue = late

The prefix mal- is very common in Esperanto and often forms the opposite of a word.

Examples:

• bona = good → malbona = bad
• granda = big → malgranda = small
• frue = early → malfrue = late

Because kiam and kiom malfrue do not mean the same thing.

• kiam li alvenos = when he will arrive
• kiom malfrue li alvenos = how late he will arrive

The first asks for a time.
The second asks for the degree of lateness relative to some expected time.

So if the meaning is specifically how late, then kiom malfrue is the right choice.

Because the arrival has not happened yet from the speaker’s point of view.

In Esperanto:

• -as = present
• -is = past
• -os = future

So:

• alvenas = arrives / is arriving
• alvenis = arrived
• alvenos = will arrive

Since the speaker says I don’t know how late he will arrive, the future ending -os is exactly what you would expect.

Veni means to come.
Alveni means to arrive or to come to a destination.

The prefix al- often adds the idea of toward or to.

So:

• li venos = he will come
• li alvenos = he will arrive

In this sentence, alvenos is more precise, because the idea is not just movement, but reaching the destination.

Esperanto does not usually do the kind of subject-verb inversion that English does in questions.

In English, we say:

• How late will he arrive?

In Esperanto, the clause keeps normal order:

• kiom malfrue li alvenos

That is very common in embedded questions and subordinate clauses. The subject li stays before the verb alvenos.

So Esperanto is more regular here than English.

Because kiom malfrue li alvenos is a subordinate clause, and Esperanto normally separates subordinate clauses from the main clause with a comma.

So:

• Mi ne scias, kiom malfrue li alvenos.

That comma is standard and recommended.

Yes, but it is an embedded question, not a direct question.

The whole sentence is a statement:

• Mi ne scias... = I don’t know...

Inside that statement, there is an indirect question:

• kiom malfrue li alvenos = how late he will arrive

So Esperanto uses kiom to introduce the content of what is not known.

Because kiom already introduces the subordinate clause by itself.

In English, we sometimes use words like that, whether, or a question word. In Esperanto, if the clause already starts with a question word like kiom, nothing extra is needed.

So:

• Mi ne scias, kiom malfrue li alvenos.

That is complete and natural as it stands.

Normally, no. Esperanto usually states the subject explicitly.

So:

• kiom malfrue li alvenos = correct

Leaving out li would usually sound incomplete, because the verb ending -os shows tense but not person strongly enough to make the subject unnecessary in normal usage.

Esperanto is not like Spanish or Italian in this respect; subject pronouns are generally kept.

Yes. That is a very good way to understand it.

A fairly literal breakdown is:

• kiom = how much / to what degree
• malfrue = late

So kiom malfrue = how late

It may help to remember that Esperanto often builds meanings very transparently from smaller parts.