Template inhibition of the qPCR can indeed be affecting your first point.  Yes, start your standard curves at the point where inhibition lets up, and then, for each sample, dilute them to a good point within the standard curve range that is not exhaustive of your samples - depending on how many targets you are looking for.  If only looking for 1 to 3 targets, you most likely have enough sample to accommodate that, and thus dilute them to the 1st good point in your standard curve - to allow the targets the best chance to bark if indeed the dog's you seek are present.

Attached is the standard curve I am getting for 1 of my genes. While my efficiency is decent (~93%) with R2 of 0.993, it looks like my first (undiluted) point is not falling in the standard curve. Every dilution point after the first one is 1:5.

Mark, to answer your question, I am trying to amplify parasite using real-time PCR, starting with cDNA

So, when using my samples for this target, does it make sense to dilute my sample 1:5, or should I continue optimizing? Also, is it possible to dilute my actual samples 1:2 instead, to avoid diluting it too much?

Your first point is fine... keep it.  Not inhibited.

To look for inhibition try this:

1. Omit the higest concentration standard from the graph, then extrapolate the graph to cross the standard you omitted (i.e. "predict" where the highest conc standard should be).

2. Calculate the mean Ct value for the highest concentration standard.

3. If the mean Ct value obtained is more than 0.5 Ct values from the predicted value from the extrapolated graph then you have inhibition.

See Annex 2 from http://gmo-crl.jrc.ec.europa.eu/doc/ENGL%20MV%20WG%20Report%20July%202011.pdf