Automotive CAN Signal Dataset Generator

About OpenTrain

OpenTrain is the #1 platform for people building careers in AI training and data labeling. We connect skilled contributors with practical projects that shape how real-world AI systems behave.

This project is run by OpenTrain AI and is part of our work helping engineers and data teams create high-quality training datasets for automotive applications. You will contract directly with OpenTrain AI.

Why AI training work matters

AI training (also called data labeling or human feedback work) is the human side of building machine intelligence. Contributors create the examples and mappings modern models learn from—tasks like annotating text, images, audio, or structured vehicle signals.

These projects are often fully remote and flexible, and they let subject experts directly shape systems used by OEMs, insurers, and product teams.

The role

You will generate a dataset that maps short, natural-language use-case descriptions to the specific CAN signals that should be collected to address each use case.

Work uses OpenTrain's internal proprietary tooling to produce the dataset in document form. This is a fixed-price contractor engagement: total pay is $2,000 USD for the completed deliverable.

• Employment type: Contractor (fixed-price deliverable).
• Data type: Document; Label type: DATA_COLLECTION.
• Tooling: Internal proprietary tooling (training and access provided).

What you'll do

Translate natural-language prompts describing vehicle behavior, feature usage, or diagnostic goals into an ordered list of CAN signals that should be collected to satisfy the use case.

Follow provided annotation guidelines and use the internal tool to record mappings, examples, and any required metadata for each item in the dataset.

• Example mapping: "I need to understand how driver uses the sunroof" -> [signal_for_sunroof, signal_for_driving, signal_for_speed, signal_for_ac].
• Ensure lists are complete, relevant, and actionable for downstream data collection and analysis teams.
• Flag ambiguous or underspecified prompts and provide clarifying notes per guidelines.

Requirements (mandatory)

Subject-matter expertise in automotive CAN signals and in-vehicle data is mandatory. Applicants must demonstrably understand how vehicle signals relate to vehicle functions and diagnostics.

This listing expects an intermediate level of experience; you will not be asked to perform software development work, but you must provide accurate, domain-correct signal mappings.

• Proven knowledge of CAN bus signals and common signal names/meanings.
• Ability to interpret use-case descriptions and map them to relevant signals.
• Comfort using an internal web-based annotation tool (training provided).
• Contractor status is required; this is a fixed-price engagement.

Preferred experience

The following backgrounds are preferred but not strictly required. List any relevant prior work when you apply so we can evaluate fit quickly.

Preference is given to people who have worked directly with vehicle diagnostics, insurance telematics, remote vehicle support, or OEM data teams.

• Experience building or using vehicle diagnostics systems that identify which signals matter for troubleshooting.
• Work with vehicle insurance teams using driving data for risk or premium calculations.
• Experience supporting remote vehicle diagnosis via signal telemetry.
• Collaborated with OEM data science teams on signal selection or feature engineering.

How the project works & how to apply

This is a worldwide, remote contractor opportunity. OpenTrain will provide access to the internal tooling, annotation guidelines, and a sample set of use-case prompts. You will complete the dataset according to those guidelines and submit it for review.

Compensation is a fixed price of $2,000 USD for the finished dataset. Time requirements are set per-contract and will be agreed before work starts; please describe your availability and estimated turnaround when you apply.

• Apply with a brief summary of your CAN signal experience and examples of relevant past work.
• Be prepared to demonstrate domain expertise during a short screening review.
• Deliverable acceptance is based on adherence to the annotation guidelines and dataset completeness.