CAN / Automotive Reverse Engineering Engineer

Part-time / Student / Junior–Mid | Automotive CAN | Reverse Engineering | ADAS

We are building an AI-assisted system that automatically recognizes CAN signals and significantly reduces the time required to add new vehicles to openpilot.

As part of this project, we are developing an online version of Cabana and training machine learning models on real-world CAN logs.
Your work will directly define the ground truth that the AI learns from.

This is a hands-on reverse engineering role focused on real vehicles and real data — not simulations or theory.


What you will do

  • Capture and analyze raw CAN logs from vehicles
  • Reverse engineer CAN messages and signals (speed, steering, brake, gear, blinkers, cruise, etc.)
  • Identify counters, checksums, multiplexed signals and message structures
  • Create and maintain DBC definitions
  • Port new vehicles to openpilot
  • Prepare labeled datasets for ML training
  • Design heuristics and validation rules for automatic signal detection
  • Work closely with the ML and backend engineers

Requirements

  • Practical knowledge of CAN bus (analysis, not only theory)
  • Experience with CAN sniffing/logging tools (Cabana, SavvyCAN, or similar)
  • Ability to read raw frames and manually extract signals
  • Basic Python for scripting and data analysis
  • Strong analytical and problem-solving skills
  • Comfortable working with messy real-world data

Nice to have (not required)

  • Experience with openpilot or similar ADAS systems
  • UDS / ISO-TP diagnostics
  • Embedded or automotive electronics projects
  • Reverse engineering or car hacking experience
  • Linux environment

Who this role is for

  • Electronics / robotics / embedded / automotive students
  • Engineers who enjoy reverse engineering and low-level systems
  • Hobbyists who like analyzing vehicle networks

If you like figuring out what unknown bytes mean and discovering how cars really communicate, this role fits you.


What we offer

  • Work with real vehicles and real CAN data
  • Build tools that are immediately used in production
  • Direct impact on an AI system that automates vehicle porting
  • Flexible hours (student-friendly)
  • Small engineering team, no corporate overhead