Don't expect unfinished projects from me.
Do you really want to know the history from the Creation of the World?
Well... I have seen it all - good projects, bad projects and everything in between. I have seen mainframes from the "good old days" when 4K RAM was more than enough to land on the Moon, and I've seen Jetson(-s) from NVIDIA (2070 FP4 TFLOPS) that aren't good enough translate Simplified Chinese correctly.
I tried to be focused on embedded software as much as possible, but I had "to eat" whatever project life threw at me.
I started looking at ML when it was just a simple project for a simple "Expert System" (let's say it was a simple linear regression without activation functions), and I missed out on the opportunity to join modern machine learning practices while not having the opportunity to do real live machine learning projects. Still Reinforcement learning is still the best :) although I think it's a straightforward way to bring machine learning into the embedded world. I saw fascinating results in robotic and I can understand what is behind these results. I really want to jump on that train... but I'll eat what life brings to my table.
Work Terms
I can work solo or as a part of a team.
I prefer 8 hour a day, with reasonable overtimes.
Payment: depends on a project.
Short projects are paid upon code delivery.
Longer projects should be divided into clearly defined stages, with payment for each stage being paid upon code delivery.
Communication (English): verbal, written. Best idea is to communicate through some PM software like Jenkins, MS Project; some bug tracking software like Jira, Bugzilla; also it could be GitHub, GitLab...
For longer projects I will insist on having clear written documentation: SRS (BRD is also good to have) before the project starts, or I will insist on developing such documentation as part of the project.