In the last three interviews (Bay Area, last month), I was asked to own perception (YOLOv8 on Jetson Orin), motion planning in ROS 2/MoveIt, and also stand up CI/CD and data pipelines for model retraining — automation + MLOps in one seat… Is this the new normal, and if so, how are you framing scope/comp for roles that span PLC integration and ML deployment?
Feels like the new normal at Bay Area Series A–C, but I only take it if comp reflects two hats and there’s a written RACI splitting ROS 2/MoveIt from CI/CD. For YOLOv8 on Jetson Orin, push them to standardize on Isaac ROS/Triton (Isaac ROS (Robot Operating System) | NVIDIA Developer) so retraining and serving aren’t bespoke. @hiring-managers, please level it as Staff/Lead and price it accordingly.
I’d frame it explicitly as a split role: about 60% ROS 2/MoveIt/PLC and about 40% ML systems (YOLOv8 on Orin + retraining), comp pegged to the higher band using levels.fyi as the anchor: https://www.levels.fyi. One concrete ask that’s worked for me is a written time allocation plus a hiring trigger (e.g., if labeling volume > N hrs/week or CI incidents > X/month, they open a dedicated MLOps headcount). Otherwise you’re the pit crew and the driver — what ranges have you been quoted for this bundle?
Ask for written SLOs and distinct on-call for ‘Jetson Orin’ retraining pipelines; did they budget labeling?