Reverse-Engineering a Car's CAN Bus with Claude
Letting an AI agent drive the hunt for an unknown CAN signal — over an MCP server, with the analysis done by real tools and the reasoning done by Claude.
Letting an AI agent drive the hunt for an unknown CAN signal — over an MCP server, with the analysis done by real tools and the reasoning done by Claude.
A dependency-free C port of Chrome's offline dinosaur game — no stdlib, no heap, 1264 bytes of state — with C++, Rust and Python bindings, and a live WebAssembly build you can play right here in the browser.
A cross-platform desktop tool for CAN, CAN FD and LIN buses: live monitoring, recording and replay, signal databases, reverse engineering, UDS/J1939/CANopen diagnostics, ECU flashing, simulation and scriptable test automation — with any python-can adapter, or no hardware at all.
A pytest plugin for the tests every model and training loop needs: inference contracts, invariance, deployment parity, checkpoint resume, gradient accumulation, dataset metrics with confidence intervals, and zero-code testing of any model.
Build custom objectives with tensor operations, reusable modules, quantile regression, composite losses, and gradient checks.
A tour of every documented core PyTorch loss, with equations, input conventions, examples, and practical advice.
Yearly book recommendations of 2025
A curated list of solo jazz moves in alphabetical order to memorize
University courses taken at the faculty of mathematics
Yearly book recommendations of 2023