Sensor Configuration Faults

Hashlist; Detroit (MI); Hybrid; USD 96,000 - 124,000; Part time

Job summary

Hashlist is seeking a hands-on engineer to build a benchmark environment around a real sensor product (radar, camera, IMU, ultrasonic or similar). You will assemble data, datasheets, mounting and calibration specs, and both good and faulty recordings.

Choose and use the required tools—CLI, compiler, QEMU or a simulator, Python, PDF tools—and design tasks that require solving via code, data, and datasheets. This is a freelance, hybrid role with hourly compensation.

Qualifications

• Hands-on engineering experience with a real sensor product (radar, camera, IMU, ultrasonic or similar).
• Experience reading sensor/SoC datasheets, mounting and calibration specs.

Responsibilities

• Build a benchmark environment around a real sensor product.
• Create tasks tied to real faults and root-cause analysis across code, data and datasheets.
• Provide correct answer, scoring script, and plausible naive answer.
• Run models to validate low scoring against your answer.

Skills

Sensor hardware knowledge

Signal processing

Python scripting

QEMU/simulator usage

Command line proficiency

Firmware reading

Data sheet parsing

Root cause analysis

Tools

Python

QEMU

Compilers

CLI tools

PDF tooling

Job description

• Working model: Hybrid
• Employment type: Freelance – hourly rate: $70–90 (paid per accepted task)
• Build a benchmark environment around a real sensor product you have worked on – radar, camera, IMU, ultrasonic or similar
• Put together the environment with real firmware or processing code, sensor and SoC supplier datasheets, mounting and calibration specs, and both good and faulty field or bench recordings
• Choose the tools the work needs – command line, compiler, QEMU or simulator, Python, PDF tools
• Create five or more tasks, each tied to a real complaint or failure on that sensor (e.g. far objects missing after a bracket change, sign flips in the suspension acceleration band, ghost targets on returned rear corner radars)
• Write tasks that can only be solved by working through the code, data and datasheets to find the root cause and fix
• Design tasks where current AI models fail – a textbook or plausible diagnosis, a missed constraint buried in the datasheet or recording, or a missed real root cause
• Supply the correct answer, a hidden marking script that scores 0.0 to 1.0, and a plausible naive answer that scores low
• Run a preset range of models against each task to confirm that the models score low while your answer scores 1.0
• Hands-on engineering experience on a real sensor product – radar, camera, IMU, ultrasonic or similar
• Working knowledge of sensor firmware or signal processing code
• Able to read and apply sensor and SoC supplier datasheets, mounting and calibration specs
• Experience finding the root cause of real field or bench failures from recordings
• Comfortable with the command line, compilers, QEMU or simulators, and Python
• Python skills to write automated marking scripts

We can only accept candidates that are willing to work as individual contractors / freelancers.