Back to the board
HashlistVerified· Posted 15d ago

Driver and Peripheral Bring-up

Type
Per-task
Location
Remote
Applying through our link may earn us a small commission — at no extra cost to you.

About this role

Your job is to find a real-world AI model failure within driver & peripheral bring ups. You will give the prompt, context, files etc. and run a preset range of models against the problem.

If you have several model failures in mind, you can submit several ones (as many as you want).

We will use your submitted data to benchmark a wide range of models, and measure how well they perform at embedded engineering tasks. For the brief:

1. Build an environment around a microcontroller or SoC (system on chip) peripheral you have brought up: CAN (Controller Area Network) controller, ADC (analogue-to-digital converter), SPI (Serial Peripheral Interface) or I2C (Inter-Integrated Circuit) device, timer or DMA (direct memory access).

  • Contents: the reference manual chapters and errata sheet, the device datasheet, the vendor SDK (software development kit) at two release versions, the driver code, and a QEMU (Quick Emulator) board model or recorded register access logs.
  • Tools: cross-compiler, QEMU, git, Python, PDF tools.
  • Put in everything a developer on the project would have open. If a task depends on a file, the file has to be in the environment.
  • Once saved, the environment appears under My environments on DOP. Write all your bring-up tasks inside it.

2. Write tasks inside the environment. Five or more is ideal, and one is enough to start.

  • Each task is a question.md plus the files it needs, built on a bring-up or SDK migration problem.
  • The model writes or fixes the driver and writes answer.json with the register values and the order it sets them in.
  • Your marking script builds the code, runs it on QEMU, compares the register write sequence with your reference trace, and checks the delays between writes.
  • A strong task turns on an errata entry or a change between the two SDK versions.
  • Supply your answer (it must score 1.0), the hidden marking script (0.0 to 1.0), and a naive answer that a competent engineer might give first, which must score below 0.3.

3. Example tasks

  • ADC readings drift after moving to the next SDK release. SPI flash writes corrupt one page in a thousand because of an errata item. CAN controller drops frames after waking from low-power mode.

4. Payment

  • $300 per accepted task. A task is accepted when it passes review,.
About Hashlist

A credible, embedded-engineering-focused AI-data platform with genuine paid briefs ($300 per accepted submission) and a generous referral. It's a young, niche product with little public review history, so confirm payout terms as you go.

AITrainerGigs aggregates this listing from Hashlist.