Hands-on work on the probabilistic robotics stack: filtering, SLAM, learning-based perception, and the plots and diagnostics you need to actually debug all of that.
- Quaternion-based UKF for 3D orientation estimation from IMU data
- Particle-filter SLAM with occupancy grids
- NeRF-based scene reconstruction
- Policy optimization for robot control
- Built the visualization pipeline (plots, GIFs) that made every step debuggable
Repo: https://github.com/cedrichld/learning_in_robotics
