I mentioned washing machine heaters because that's what I use and I can find them cheap, a bucket of water usually gives me plenty of time at the power I need to test.
The first setup I had was thick nichrome wire drowned in brick-sized concrete blocks, those got hot too fast.
If long test times are a problem you'll have to consider overall dissipation, a large thermal mass is only convenient up to a certain point. Once size and cooling times start getting in the way you may as well go for something with forced air flow.
If you're on a budget there are other things that could be used such as hair dryers, toasters, fan heaters... That may work out cheaper than power resistors.
For current sense I'd use a shunt resistor (+opto) it makes the load easier to use on DC if needed.
The PWM filtering / snubbering will depend on the layout, the other aspect is filtering for control loop stability.
You may also want to measure input voltage (line losses) if a constant power mode is needed.
Another approach would be once more using a resistive load to set maximum power consumption and a triac with phase angle control, closed loop can be added. Definitely dirty and AC only.