It is certainly not true that "all male humans have roughly 2.5 (or whatever the number) kids". First of all, male and female humans have exactly the same reproductive rate. For obvious reasons, every time a male has offspring, a female must have had also. Last I checked neither male nor female humans are capable of parthenogenesis (certain popular religious beliefs notwithstanding).
Second, lets assume that the 2.5 number is correct. That would be the average number of children per couple. That does not mean that all couples will have 2.5, or even that most couples will have 2.5. It just means that the average will be 2.5. If, for example you have one couple with 6 children, one with 2 and two with 1, the average will be (6+2+1+1)/(1+1+2)= 2.5.
On to the main point. What does selection mean? In its simplest form, that the individual most likely to survive (the famous "fittest") is also most likely to reproduce. This is a very simple concept, the longer you live the higher your chances of managing to have offspring. If you die two weeks after birth it is going to be hard to manage to reproduce yourself. This has not changed.
So, what does "fitness" mean? It can mean many things. If you are a warm blooded creature at the beginning of an ice age for example, it could mean being better at regulating your temperature than your peers. If you are a 21st century human, it could mean being funnier on twitter than your peers. The two are not fundamentally different. They are both selection. As long as one mate is chosen over another selection is happening and the "fittest" (in each particular context) is most often selected.
Add in the process of culturally modified selection pressure, and it
seems to me that even an "unfit" male would end up having a couple of
offspring. The fittest male (or female) is no better off than his or
her contemporaries because of this "leveling" effect.
"Culturally modified selection pressure" as you call it, is still selection pressure. Cultural factors can change what it means to be "the fittest" but there is no objective gold standard of "fitness". While it may be true that in modern human society, different characteristics are selected for than was the case with early Homo sapiens, this does not mean that "evolution is not occurring". On the contrary, it is occurring but perhaps it is moving in a new direction. In fact, this is essentially a circular argument. By definition, "fittest" means most likely to survive and reproduce. It does not mean strongest or fastest or prettiest. It just means whoever is better at reproducing. If that happens to be those individuals who are best at square dancing, then it is they who are the fittest.
Take the example of a modern human with diabetes. Medicine allows diabetics to lead fully productive and largely normal lives. So, perhaps diabetes is no longer a selective criterion. This does not mean that the diabetic cannot be selected for or against based on their fitness on other scales.
Whatever the selective pressure, whatever it may be that defines a "good mate", if selection is present then so is evolution. The only way to remove a species from the process of selection would be to have all (or none) individuals of each and every generation reproducing at the same rate. This is clearly not the case with humans. Surely not everyone around you has, or will have, children? There you go, selection!
In answer to your comment, yes indeed, in order for a selective pressure to make itself felt and affect phenotype (at the species level), it needs to be constant across several generations. However, even the absence of selective pressure affects evolution. As others have mentioned below, active selection is not the only mechanism of evolution.
Your main question however seems to be the following: If modern society (medicine etc) allows individuals that would not survive in the wild to reproduce, how does that affect evolution? The main points in my answer, and all others here, are:
Even if we accept that modern humans have removed themselves from the purely "biological fitness"-based selection pressure (an assumption I am not at all sure is true), and assuming that this removal is constant enough over many generations (again unclear), even if all this is true, evolution is most certainly still occurring. It may even be faster since genotypes that would not survive in the wild persist in the gene pool, thereby increasing its diversity.
As you point out in your comment below, for such social pressure to make itself felt, it needs to be constant across many generations. We are probably not there yet.
Most importantly, as I said above, there is no such thing as an absolute biological fitness. When the ecosystem changes, so does the definition of fitness. Modern humanity's ecosystem, our habitat, is intimately connected with our culture and society. If an individual is better at reproducing in that context, then that individual is more fit.