Any advice on prep for Perception/Planning/Control interviews at companies like Waymo and Cruise? In addition to: - Google/leetcode style questions, especially ones that are less combinatorial/string-based and are more geometric or spatial. - Google style design questions - Review machine learning (basics like SVM, linear through stuff like sequence to sequence and self attention models) - Review linear algebra (perspective math, spectral theorems, exterior algebras, canonical forms) Any other suggestions? Thanks!
Can you share your previous work experience related to the position?
All of the above. Personally - consider Aurora.
Thanks. I'll check Aurora out.
What is Aurora?
You should know what is a PID controller (and other topics in control theory, like linear control theory, Laplace transform, Fourier transform, stochastic control, fuzzy logic, etc.), Kalman filtering (and other topics in stochastic system modeling like particle filtering), occupancy grids, some machine learning (perceptrons, convolutional networks, etc.), graph algorithms (especially ones that work with real world 2D/3D topologies like A*), etc.
Ditto
Based on my interviewing at Lyft, Baidu, Didi, Uber, Argo (and have Cruise on-site soon): linear algebra, linear-non linear optimization, 3D geometry (point clouds), transformations, basics of kinematics, computer vision fundamentals + multi-view geometry details (if that’s your specialization), probability basics, machine learning basics, deep learning basics (more details if you are a deep learning guy). Exposure into path planning (graph search techniques), control sydtem/robotics will also help. At Lyft, Uber, Baidu, Didi, the software questions were typical Google style (similar difficulty); probably it was slightly easier at Argo.
Hey OP, did you land the job or still looking?
How did your interviews go? Did they ask any topics other than discussed here?
++
Interviews went well! Everything I was asked was related to vehicles. I didn't get asked anything that I prepared for, although Leetcode certainly helped for some parts. Not to divulge specifics but imagine stuff like: lanes and passing, interpolating points in a lane of traffic, finding a path in space that avoids obstacles, etc.
Tech Industry
8h
902
I haven’t done shit today!
Tech Industry
12h
2834
Avoid teams with only Chinese or Indians especially with a Chinese/Indian manager
Tech Industry
Yesterday
38373
Worried that our top performer is an attrition risk. How do managers handle this?
Personal Finance
Yesterday
3058
Dave Ramsey advises against 529
AMA
Yesterday
3752
I’m a professional coaster AMA
Wouldn't doubt you get some computer vision questions in terms of processing.