Foundations

You don't need a PhD to build a robot. You need three things — and you already know two of them. A practical map of the math you'll actually encounter.

~12 min

NumPy, SciPy, PyTorch/JAX, matplotlib — four libraries cover 90% of robotics Python. Here's the minimum fluency and the idioms that actually show up in robot code.

~14 min

Shell fluency, ssh, tmux, systemd, udev — the half-dozen tools that separate people who kind-of-know-Linux from people who can actually run a robot in the field.

~16 min

Versioning code is solved. Versioning the URDF, the firmware, the CAD, and the BOM together — that's the part that bites every hobbyist. Here's the layout that scales.

~12 min

Docker is the standard way to ship a robot's software stack in 2026. The 20 lines of Dockerfile that get ROS 2, your code, and your dependencies running anywhere — laptop, robot, CI.

~14 min

Vectors, matrices, rotations, and eigenvectors — every example a robot. The calculator-level fluency you need for kinematics, dynamics, control, and SLAM.

~18 min

Bayes, Gaussians, and the covariance matrix. The math behind every Kalman filter, particle filter, and SLAM paper — distilled to the parts a roboticist actually uses.

~16 min

Only what robots need. Derivatives as rates, gradients as steepest-ascent, Jacobians as the matrix that connects joint velocities to end-effector velocities — and almost no symbolic integration.

~13 min

The silent source of half of all robotics bugs. Name your frames, declare your units, watch your handedness — discipline that costs ten minutes and saves weeks.

~12 min

Robotics papers are dense and prone to fluff. Here's the read order, the skip-list, and the questions that separate breakthrough from bluster — so you spend an hour per paper instead of three.

~11 min