HowToBeABetterRoboticist

“How to be a Better Roboticist”, from Jim Teza at CMU Robotics Institute/Field Robotics Center. Really I think this applies to being a better engineer in general, software or hardware, though there are a few robot-specific wrinkles.

Never be satisfied and always seek perfection, but also know when a robot is ready to test.
Think about your work. If you did something incorrectly, fix it.
Become an expert in the domain your system is for. If you’re building a nuclear robot, learn all you can about nuclear. If you’re building a moon robot, learn all you can about the moon and how to get to the moon. etc.
Never stop learning. Keep learning about new advances in your field and all the surrounding fields.
If there is an issue, express frustration, then move on and solve it.
Build people. Spend the time to teach and show others how to correctly do things. It will pay back with dividends.
Spend the time to debug and solve the little issues. Those little issues have a way of coming back in the field.
When debugging always distrust your tools. Your tools might be giving you bad data that leads to misdiagnosing the problem with the robot.
Distrust all hardware and assume things won’t work properly at first. Then be surprised when things actually work. This also helps since you plan for contencies in your designs, schedule, and field deployments.
Know when to fix things yourself, and when to go back to the person responsible.
If a component is suspect, label it “bad???” or throw it away. Don’t let others use that component unknowingly. How many times have you used a bad Ethernet cable that someone else knew was suspect?
It does not matter how senior you are, there is no shame in asking another person to verify your design and work.

I notice a few themes here:

Fix broken windows
Invest in people and teamwork, including yourself
Make very sure you have the details correct, because that will carry on throughout the whole system.
– but there’s a balance too, be the right level of perfectionist.