One operational point is to leave a surge buffer at the top of the tank, based on maximum possible input flow. I would eliminate the control valve, to use the VFD, but you may need a stop valve if the discharge of the pump allows gravity flow, when the pump is stopped for Low-Low Level, for instance. A stop valve is easier to apply and more reliable than a vacuum breaker on the discharge line.
The VFD provides significant energy savings, but perhaps as valuable is the reduction of wear and tear on your pump, especially centrifugal. The wear increases according to the square of the pump shaft speed. The reliability of inverter equipment is much better than it was 20 years ago.
Starting & stopping a motor will require a small analysis of the heat load accumulated in the motor rotor & windings. The allowable watt-seconds is directly related to the motor size, where a small 1/2kW motor might handle even indexing service, where the same mechanical load & cycle with a 2kW motor might burn up the windings. You need to evaluate what the worst case cycling will be, and set up your level set points to remove rotor heating as a problem.
One more thing I forgot to mention, for a VFD application with a centrifugal pump. For a given hydraulic condition, there is a minimum pump speed that will produce flow. For installations where the process allows it, I usually set the pump minimum speed where it always produces a flow, regardless of tank level or discharge conditions. I set it in the VFD hardware settings, so no one at a control screen can accidentally disable it, for plants with uncontrolled operating environments.
The other advantage of setting minimum speed at the VFD is when the VFD is put into RUN mode, the pump shaft always turns, even with a speed reference equal to zero. This make LOTO and personal safety procedures easier and more reliable to verify.