I work in industry and some interesting idea has struck me. This is a conjecture which will take a year or two of full time research work to confirm it although it looks very realistic. Unfortunately I cannot apply for post doctorate or a grant due to my work commitments but I would be glad to publish an article if a reputable CS journals would be interested. This may solve very hard practical problem but surprisingly it is difficult to monetize so I cannot do it in industry.
You can certainly publish an article based on a conjecture. (One famous case is Subhash Khot's 2002 paper introducing the Unique Games Conjecture.) On the other hand, you need to include some supporting information in the paper. Why is your conjecture plausible? Can you supply examples or applications? What are its implications or connections with related work? Answering these questions may be much easier than resolving the conjecture, but they can still require some serious work.
You're most likely to be able to publish the conjecture if there's no straightforward way to resolve it. If it's clear in principle what you should try, and the only issue is that you don't have time to try it, then you're more likely to meet resistance (it may still be possible, depending on the publication venue, but this would probably rule out many of the more prestigious conferences).