The best way, I think, at least, to understand quantum field theories is to start with a very simple one: phonons in a crystal lattice. In some sense, this is actually the only "really fully understood" kind of quantum field theory, because as soon as you talk about interacting quantum fields, the mathematics in continuous space no longer provides for a unique solution determined by the behavior over a limited range of scales.
In this system, we have an infinitely long chain or train of harmonic oscillators, notionally representing atoms in a one-dimensional crystal (like a metal like iron), at uniform spacing. Importantly, the oscillators are also coupled to each other, in that the Hooke's law restoring forces are provided by connections - which you can think of as like springs - to their adjacent oscillators.
And what you can find, then, when you work through the maths of this system, is that you can set up a quantum state for the oscillator chain that looks like a single oscillator has been "plucked", i.e. excited to its next higher vibrational level, while all the other oscillators remain at ground state (this state is not stable, of course, any more than its classical analog, i.e. a string of balls with springs between them where you've carefully stretched one ball out of equilibrium while holding the rest in place). That is, for each lattice position $i$ (where we imagine the atoms in the chain as numbered with integers), we can consider there to be a quantum state of the lattice $|i \uparrow \rangle$ where that that particular oscillator has been promoted to its first excited level, while the oscillators around it are still at ground level. And because these are separate and orthogonal quantum states for each index $i$, we can form nontrivial quantum superpositions. In particular, we can write a general state of the form
$$|\psi\rangle = \sum_{i=-\infty}^{\infty} \psi_i(i)\ |i\uparrow\rangle$$
with weighting factors $\psi_i(i)$. These weighting factors are suggestively labeled, because what this state suggests is that we have a "pluck" of a single oscillator, since that's what each $|i \uparrow \rangle$ going into the sum represents, yet with indeterminate position, given by the probability amplitudes encoded in $\psi_i$, that "fuzzify" (reduce information about) just which of the states is "meant", and thus fuzzify which lattice position is involved in the "plucking".
And that sounds an awful lot like a particle! It's the tiniest, most local excitation possible, but it is simultaneously fuzzed out with regards to where it is, in the same manner that an electron is a point-sized object, yet "where it is" is similarly ill-determined.
And we call that a phonon.
If we expand the lattice to two dimensions, so we can do something like the double-slit experiment, and then evolve the state, we will find the behavior is that $\psi_i$ spreads, just like the wave function of an electron does, and it is that, that interferes with itself.
So what is interfering with itself is the positional wave function of the phonon, exactly like how the electron's positional wave function interferes. And since the maths for electrons is almost entirely analogous if we ignore interactions at the field level, we can say that what is happening in that case is the same thing. The electron in QFT language is a point excitation of the Dirac field, that is then "fuzzed out" as to just which point is excited via a nontrivial positional wave function, which is also the electron's positional wave function. When it goes through the slits, that wave function interferes.