I have been tasked to port a project of porting a ML project that has been running on someone's laptop to the cloud and make it production ready. It is very difficult for this person to give the full picture of the process.
This could be due to he is not actually remembering very well, given that the project has been actually done 1 year ago, or he doesn't like to lose the single point of failure / control situation.
I actually don't care about the reason and I want to get this project to success.
How may I deal with this situation?
Current blocker is that I am unable to properly estimate the time it may take for this project.
Are you trying to get the info in a meeting or asynchronously? I'd think about using both to make more progress with this knowledge transfer.
You could create a template of the information you need and start filling it out to the best of your knowledge. You can then rely on tried-and-true human behavior: people like to correct others more than they like to propose an idea. Your attempt at filling in the details may spur him to start contributing.
Can you set up a meeting with them where you can at least get their code to run on your own laptop?
Hi everyone,
Thank you very much for your feedback. I like these pointers. Since my last comment, I have made some progress.
I managed to setup the code and run it on my local machine, which helped me to understand some aspects of the code. Furthermore, I worked on a possible architecture and involved him in the discussion.
These steps helped me to progress. It is not done yet, but I am making significant progress now.
Thank you.
This seems like a pretty straightforward task to me: You need to get their code onto GitHub and attach some sort of README to it so people can understand how to run/extend it.
As Charlie mentioned, I think just having a meeting is best. Frame it as a collaborative exercise:
The important thing (as always) is to communicate with empathy, which I break down more in my Effective Communication series: Alex's Guide To Effective Communication