Let $\beta$ be any number with $|\beta| < 1$. Notice
$$\begin{align}
\sum_{n=0}^\infty \frac{\beta^n}{n!} \int_0^y x^n e^{x} dx
&= \int_0^y e^{(1+\beta)x} dx
= \frac{1}{1+\beta}\left(e^{(1+\beta)y} - 1\right)\\
&= \left[\sum_{p=0}^\infty(-1)^p\beta^p\right]\left[e^y\left(\sum_{q=0}^\infty \frac{\beta^q}{q!}y^q\right)-1\right]
\end{align}
$$
By comparing coefficients of $\beta^n$ on both sides, we get
$$\int_0^y x^n e^x dx = n! \left[ e^y \left( \sum_{q=0}^n \frac{(-1)^{n-q}}{q!}y^q\right) - (-1)^n\right]\tag{*1}$$
Ideas behind such a closed form
To shine some light why the closed from looks like this, let us re-derive the integral in a completely bogus manner. It is known that for any smooth function $f(x)$ over $\mathbb{R}$, we have
$$\frac{d}{dx}\Big[ f(x) e^x \Big] = ( f(x) + f'(x) ) e^x\tag{*2}$$
If we view $\displaystyle\;D = \frac{d}{dx}\;$ as some sort of operator over the space of smooth functions. $(*2)$ can be reinterpreted as
$$ L^{-1} D L = D + 1$$
where $L$ is the operator of pointwise multiplication a function by $e^x$. In certain sense,
integration is the inverse of differentiation. Expression of the form
$$e^{-y} \int^y e^{x} f(x) dx\tag{*3}$$
can be interpreted as applying the operator
$$L^{-1} D^{-1} L = ( L^{-1} D L )^{-1} = (1 + D)^{-1}$$
to the function $f$. If we close our eyes and expand
$$(1 + D)^{-1} = 1 - D + D^2 + D^3 - \ldots...$$
We get a formal expression for the integral
$$\int^y f(x) e^x dx = e^y ( f(y) - f'(y) + f''(y) - \ldots ) + \text{constant}\tag{*4}$$
Even though the logic is bogus, if we apply this to $f(x) = x^n$, we reproduce the formula $(*1)$.
In certain sense, $(*4)$ can be viewed as the result of repeatedly applying
integration by parts to $f(x)$. When $f(x)$ is a polynomial, the chain of integration by parts terminated in finite steps and we get back another polynomial multiplying $e^x$.
That's why $(*4)$ works when $f(x)$ is a polynomial and why $(*1)$ have such a form.