(You might also ask: If a shipping container full of solar panels, 74-series logic chips, & all necessary manuals went back in time 100/200/500/1000 years, what USE would people have for any of it)
If I had unlimited time to prepare but couldn't take anything with me but knowledge, I think I'd concentrate on mechanical computers and figure out what "modern" developments would have made machines like Babbage's Difference Engine not largely a dead-end. Are there just a few small things (I have no idea -- better bearings? a high tech lubricant? improved machining tolerances?) that would have made them downright practical? As a bonus, your better manufacturing will be good for a lot of things, from better (or any) pocket watches to better firearms, so you'll be able to make plenty of money while you concentrate on your 'computer' creation.
A similar line of argument might lead you to decide on vacuum tubes as the base for your technology. Glass blowing's been around for a long time and materials purity is probably not QUITE the problem it is for semiconductors. Batteries are pretty old inventions, too. Mechanical relays are pretty plausible too, and you can build digital circuits from them as well.
The concept of "stored-program computing" would be important to bring back, including the "subroutine". So would the concept of "assembling" a symbolic program into the computer's own representation. If you've got mathematicians handy in the past they'll catch on fast and work out data structures and all that other "CS stuff", albeit with some twist you'll be surprised by. I wouldn't worry about bringing a specific modern language like rust or python back to the past, personally.