There are essentially two ways to split files up with LaTex: using \include and using \input.
Let's start with \input. When LaTeX is reading blah.tex and comes across \input{blurb.tex}, it just pauses processing blah.tex and starts slurping up blurb.tex -- and when it's done it goes back to blah again, until it finishes. So basically you can quite simply "carve out" anything that you would put in your "main" file, and stick it into a sub-file, which is then "read in" with \input{}. That gives you one form of modularity. Its key issue is that every time it processes a document, LaTeX goes through this "reading in" process: so although your files are modular, the typesetting process really isn't: LaTeX goes through everything, every time.
Now of course that may make you think: suppose I split my document like this:
\documentclass{book}
\begin{document}
\input{chapter1.tex}
\input{chapter2.tex}
\end{document}
Now if I comment out one of those input lines, I can have a nice little copy of just Chapter 1, or just Chapter 2.
The trouble is that those chapters will often include things like cross-reference labels, or table of contents entries, or figure numbers, that you need to refer to in other places. That is where the slightly "cleverer" system of \include{filename} comes in: it allows you to save and use bits of the scaffolding that these sub-files have generated without actually typesetting them, using \includeonly{} to tell LaTeX which files to typeset. It's in some ways more limited though, and for present purposes assume that it's really only wanted in a very long document (we're talking book- or thesis-length not article length) where you not only want to work on editing only part of the files at a time, but you also want to control what gets typeset on any run.
So what your adviser is suggesting is actually a very accurate description of what you want. In a simplish project, you are probably best sticking with \input rather than using \include. A main .tex file (you can call it anything you want) is going to set up the "common parts" like a preamble that will specify the packages you want, the title, the table of contents (if you have one), perhaps the bibliography. But then the substance of each section is going to be in a separate file. Your main file will have an \input{} each of the sub-files at the place you need them to go. To typeset this, you will pdflatex main, and as LaTeX chomps through it will pause as necessary and slurp in the stuff from your sub-files.
So a minimum example would look something like:
main.tex
\documentclass{article}
\usepackage{lipsum}% and/or whatever
\title{Modular Papering}
\author{Mo Module}
\begin{document}
\maketitle
\input{sectiona.tex}
\input{sectionb.tex}
\end{document}
sectiona.tex
%NOTE: NO preamble or documentclass -- this is just being "read in"
\section{Section A}
\lipsum % Will work because I loaded the package in the preamble
sectionb.tex
%NOTE: NO preamble or documentclass -- this is just being "read in"
\section{Section B}
\lipsum
And just pdflatex main to get the complete document, with a title and two sections. All the real editing work will probably be done on those sub-files.
As has been pointed out in the comments, there are some good resources you can find by searching online. Two, pointed out in the comments, that give particularly clear explanations are https://en.wikibooks.org/wiki/LaTeX/Modular_Documents and When should I use \input vs. \include?
EDITED TO ADD: I've no idea what the "tokens" you are being told about are, but my guess is they are like "tokens" for a railway train on a single track line -- one person can edit one section while another edits another, without risk of collision! Maybe a slightly rudimentary approach to version control! But certainly one advantage of splitting things into reasonably small files is that it keeps versioning neat.
\documentclassand ending with\end{document}. It should work like this, here an example: https://tex.stackexchange.com/a/252/124842 – Bobyandbob May 17 '17 at 14:21