IBM

A lot of interesting problems for an engineer to tackle. The clients can also be very interesting.

Patent submission can be very fun and is part of the review process. If you feel that you are creative, this can be a hoot.

A friend of mine once said that no one was willing to knife the baby. There is so much pressure to be on a winning team that it's easier to tear down other groups rather than improve your own. I actually had a second-line manager denigrate our 5x speed improvement because we didn't spec it at the beginning of the project.

Find more objective ways to measure performance; the current process is way too political.

Reduce pressure for groups to affect other groups. There is value in improving one's own group, but the current management culture dismisses anything that's not outwardly visible.

Also, find some way to audit patent appraisal. The current method creates "good old boys" that approve patents based on who submitted.