I'm studing signed Lah's numbers $L(n,k) = (-1)^n\frac{n!}{k!}\binom{n-1}{k-1}$.
It's easy to show that exponential generating function for the sequence $\{L(n,k)\}_{n\geq0}$ is equal to $$ \sum\limits_{n=0}^{\infty}L(n,k)\frac{t^n}{n!} = \frac{(-1)^k}{k!}\left(\frac{t}{1+t}\right)^k. $$ So what's about ordinary generating function $\sum\limits_{n=0}^{\infty}L(n,k)x^n = ?$. I'm trying to apply connection between ordinary generating function $A(x)=\sum a_n x^n$ and exponential generating function $B(t) = \sum a_n \frac{t^n}{n!}$: $$ A(x) = \int\limits_0^{\infty}e^{-t}B(xt)dt, $$ to prove that $\sum\limits_{n=0}^{\infty}L(n,k)x^n = x\prod\limits_{k=1}^{\infty}\frac{x+k}{1-kx}$.
So how to prove it?
For fixed $k$, note that $L(n,1) = (-1)^n n!$, so for $k=1$ the power series would be $$ \sum\limits_{n} (-1)^n n! x^n. $$ The problem is, this power series doesn't converge for any real $x \neq 0$, so it can't be expressed as an analytical function. The terms grow even faster for larger $k$, so it's true for them as well.
If you want an OGF with a fixed $n$ rather than fixed $k$, you essentially get Laguerre polynomials, which are also quite far from the form that you want to prove in your question.
As I also mentioned in the comments, the RHS of your expression doesn't meaningfully converge, and if we limit the product up to $k$ rather than up to $\infty$, we get something that seems to have nothing to do with Lahs numbers. So, I don't think this question can be meaningfully answered without further clarification.