That’s right you can build one without making any changes to Corda and when I get time I’ll publish the CordaCoin design but won’t be until next year. That being said, I wouldn’t recommend building it as crypto coins don’t really solve any “real world” problems – they are a lot of effort to implement and come with too many trade-offs like block re-orgs and handling them properly, as Mike mentioned.

I guess the risk is in reducing its value for privacy. You could do something similar to what Satoshi originally proposed for "BitDNS" and embed the salt into e.g. a command which you then grind, allowing you to present the PoW independently of the data using a Merkle branch (FilteredTransaction). That would allow the privacy salt to remain fully random.

However, note that grinding the tx hash like that would require you to sign the tx after the PoW is found. 

Roger did a talk at CordaCon on "CordaCoin", a project of his that explores the integration of PoW into Corda. But it doesn't really fit much with the rest of the platform's design, which assumes finality. Re-org handling is extremely painful from both a programming and business perspective.

thanks,
-mike