I believe it does converge uniformly to 0. Note that $|f_n(x)| = \frac{nx}{1+n^2x^2}$ and its derivative is $$\frac{n-n^3x^2}{(1+n^2x^2)^2}$$ which is zero at $x = \sqrt{n/n^3}$ where the function has its maximum. So this function has a peak which moves towards $x=0$ as $n \to \infty$, which means that for some $N$, $\forall n \ge N$, the location of the peak has moved to the left of $x=a$, and since the function then decays we have $\forall n \ge N$: $$\sup_{x\in [a,\infty)}|f_n(x)|= \left. \frac{nx}{1+n^2x^2}\right\rvert_{x=a} = \frac{na}{1+n^2a^2} \to 0$$
Is my reasoning correct? Could you make it more rigorous?
You can make it more precise by noting that $\frac t {1+t^{2}}$ is decreasing in $[1,\infty)$ (because its derivative is negative there) so $f_{n} (x)$ is decreasing for $x \geq a$ as long as $na \geq 1$. So you can take $N=[\frac 1 a] +1$.