Both the vision systems at Tesla and the bot-detection at Twitter are classifiers. Both would have trained on large datasets. Both would have domain specific feature detection sitting below a more general algorithm. Both would have a similar basis for evaluation. Both have real-time constraints on the classification problem. An engineer (who is probably an ML specialist) familiar with one would not be starting from scratch in understanding and evaluating the other system.