There were some threads on GBF about DC immersion heaters a few years ago. ISTR the voltage-versus-current (or V vs Power) characteristics of solar panels are a poor match for those of resistance heaters. So when the sun shines a little less than the design point, the power absorbed by the resistance heater gets very much less.

Some folks were trying to build MPPT devices by having an immersion with several different-resistance elements and constantly switching each element in and out series/parallel to maximize the power as the sunshine varied. There was another idea to use a variable boost/buck PSU to constantly adjust the element voltage to make the best of the PV power. Or use a commercial PV battery charger and then run the immersion off the batteries.

I haven't looked recently if anyone still does this (or if you can buy one off the peg) but if you are investing in an AC inverter, why would you also want a DC immersion system? (Except for tinkering value!)

If you do DIY then sure you are aware of safety issues with highish power DC, if there are enough hundreds of Watts to heat your water then that's enough to start a fire or weld switches closed etc. The immersion must have a DC thermostat and a separate non-resetting safety cutout to prevent boiling/ scalding.

Re immersion diverter:
If you are looking for DIY fun and the highest efficiency, this is the one:
https://mk2pvrouter.co.uk/
It does output data on 433/866MHz radio link, so there is scope of piping the data into a computer and then anything is possible.

Paul Camilli fairly recently (late last year or sometime earlier this year) installed an off-grid PV inverter at a house down a track from his house (Torran schoolhouse?) which can operate without either a battery or a grid connection: that is it'll produce AC output straight from PV panels.

Obviously you can't take more AC power out than the DC the panels are producing, less inverter losses, (I assume the AC voltage would sag and the inverter would cut out if you tried) so there'd still need to be some smarts to match the load to the sunshine available but it could all be done using conventional immersions and switching AC current.

Some previous discussion of mine on the problems of directly connecting PV panels to an immersion:

https://edavies.me.uk/2012/11/pv-immersion-gotchas/

Article, published on my site, by somebody else who did it:

https://edavies.me.uk/2016/07/mp-pv-heater/

Sorry to keep linking to my own blog but these topics do keep coming up so it's convenient to have my comments available to hand.

Never tried it, but it did occur to me that someone more skilled than me, could modify a commercial MPPT battery charge controller to run a DC immersion instead.

I anticipate that device will include the high frequency buck/boost converter that bhommels mentioned, complete with meaty high-current semiconductors and heat sinks, with a MPPT controller so that it always presents the optimum impedance to the solar panels. It might have the data logging that borpin mentioned, and some protective functions.

The device would also include battery charging controls which might need to be disabled or tricked to make it supply optimal power to a purely resistive load.

In order to heat a sensible amount of water in a sensible time, it would need a lot of Watts, so the equipment would be expensive and (even if high efficiency) would give off a lot of heat. If any of the system runs over 50V (edit: or is it 120V? Or 0V?) then I think you might need Part P approval in England.

On the subject of safety, this short video popped up of someone demonstrating an arc between the tails of (I think) about 1kW of panels. https://youtu.be/S9a2oPCIMr0