No problems.  Even seasoned engineers might not learn how to use their scopes when it comes to high speed signals.
For a switch mode power supply, 100mV peak to peek noise is pretty much normal.  Published power supply noise figures (e.g. datasheet) are typically measured with a bandwidth limit filter on.  (scope menu on right)

Since you have long external wires to LED, that's probably where the parasitic inductance comes in.  Not a whole lot you can do on your board layout there, but now you know why that happens.  The next question is whether or not you need to fix it.

For personal projects, that's fine.  Now if you are trying to sell a product and have FCC/EMC concerns,  then there are things that you could improve upon.

Awesome, thanks for the info. I'm guessing that having an ESP8266 will be enough of a challenge with the FCC if I do go down that track. At the moment, the unit is working great, it hasn't reset or had any problems since I first wired it up. I figure that will be a challenge for another day. 

Best of luck with your projects. I'm really keen to see how your Sonar project goes! I see your cover picture pretty much every time I jump on hackaday.io

I thought of something... May be a snubber connected between the drain/source of the MOSFET right across the connector might slow down the slew rates to minimize those undershoots.

Thanks for the Skull BTW.

no prob for the skull. I have had a little play with the esplux this afternoon, unfortunately not enough time to do too much experimenting. I have been playing more with my scope than anything else. Thanks for the ideas. I don't have any suitable caps floating around (or snubbers for that matter) so I'll have to get some more parts in. I did however have a bit of a play with some diodes, I didn't succeed with them at all. I put one across the MOSFET, also tried across the output. No dice. I'm keen to try remove that spike still. I put my lab supply (@12vdc) on instead of the 12vac one I have at the moment with no difference, which I guess is to be expected. I'll keep playing! 

First off don't know whether or not you know how to properly use the probe for fast signals i.e. not using the big long ground lead.
http://www.analog.com/library/analogDialogue/archives/41-03/time_domain.pdf

If you have a large ground loop, then the probe might be looking at artifacts - high di/dt + inductor = big voltage spike.  In you case, it is undershoot.  There is also the case for bad grounding and you introduce a parasitic inductance into high current connection or both...

BTW if you want to look at the noise on your 3.3V  rail, there is the AC coupling on the scope - usually a button.   What it does is to show you the AC component part, so you can zoom in to 100mV or smaller and get you higher resolution to see the noise.

I highly recommend that you might want to learn to use your scope...  

Hi there, thanks for your reply. Yep, I haven't played around with it much at all yet. I'll be the first to admit that I don't know how to use it properly, which is why I have asked for peoples input. I can look around on the net all day for info, but without knowing that that is 'undershoot' I find it hard to know what to search for.

My understanding is that my probe is calibrated properly, this is the square wave from the scope itself: 

Thanks for the ground loop info, I was using the ground lead that came with the scope. I pulled that off and used the spring instead, unfortunately I still have the same undershoot thing happening. I take it it is either something to do with my board layout, my failure in understanding how to use a scope or a component not acting perfectly. Here is the 3.3v logic level out to the MOSFET:

In hindsight, I should have used AC coupling for my pictures. Here is the 3.3v rail, 8w LED load, 50% brightness one at 50mV divisions. 

Thanks again for your reply. I appreciate your feedback. I will go and read through the document you linked to.