Our group wants to perform an RNA sequencing project on a non-model eukaryotic microalgae to study it's glucose metabolism in different environmental conditions. I have very limited RNA sequencing experience and I don't know what things I should look out for and be aware of before I start. I'm wondering if differential gene expression pipeline should look different based on working with an organism with a poorly annotated genome. I've been looking at using the nf-core RNAseq pipeline that pretty much automates getting transcript abundance in the cell but I'm wondering if it is the appropriate for the job. Also is there a good, modern gene prediction software I can use to compile all the RNA and DNA sequencing data for this organism and attempt to create a more robust predicted genome annotation than already exists?
Devil is always in the details and depending on how complex the genome of the organism is (size, number of chromosomes, repetitive nature) that you are working with your level of success will vary. At this point there (should/may) be a related genome in GenBank. If so you could have a starting point to work from. More sequencing you do (from different life stages of the organism, if possible) the better represented will your transcriptome be.
Creating a good (enough) transcriptome will itself be a significant undertaking, especially if you have little or no data available in public databases. You have already received good suggestions about what to do for that below. Do use a mix of short/long read data if you can.
Unless you do whole genome DNAseq you will not have any idea about gene structure. That would be a completely different kind of sequencing. You have not indicated if you are planning to do this.