MissionRobo career roadmap
Perception / Computer Vision Engineer
Skydio · Anduril · Dedrone · Apptronik · 20 weeks
01CV foundations
required
Camera models, intrinsics, extrinsics
Pinhole + distortion, what camera calibration actually computes.
2 resources
recommended
Classical features (SIFT, ORB, FAST)
Pre-deep-learning detection. Still used inside many SLAM stacks.
1 resources
required
Epipolar geometry + triangulation
How two cameras give you depth. The math behind stereo and SfM.
2 resources
02Modern deep-learning vision
required
Deep learning for vision (Stanford CS231n)
CNNs, attention, modern training tricks.
2 resources
required
Object detection (YOLO → DETR)
YOLO for speed, transformer-based detectors for quality.
1 resources
recommended
Segmentation (Mask R-CNN, SAM 2)
Per-pixel labeling for scene understanding.
1 resources
recommended
Depth estimation (stereo + mono)
Stereo block matching, monocular depth networks (MiDaS, Depth Anything).
1 resources
optional
NeRF and 3D reconstruction
Neural radiance fields, Gaussian splatting, novel view synthesis.
1 resources
03Sensor fusion + calibration
required
Multi-sensor calibration (Kalibr)
How to align camera + IMU + LiDAR coordinate frames.
1 resources
recommended
Camera + LiDAR fusion
Project LiDAR points onto images, project image features into 3D.
1 resources
04Production deployment
recommended
TensorRT for edge inference
NVIDIA's optimizer for running models fast on Jetson.
1 resources
recommended
ONNX as the interchange format
Convert between PyTorch, TF, JAX, edge runtimes.
1 resources