← Takaisin työpaikkoihin

Control Logic and PLC Debugger

  • Etätyö
  • Ruotsi
  • Englanti
  • Julkaistu 05.10.26

Your job is to find a real-world AI model failure within PLC debugging & control logic. 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 an automated machine, line or process cell you have worked on. - Contents: the PLC (programmable logic controller) program in ladder, structured text or function blocks (IEC 61131-3 or similar), the I/O (input and output) map, device and drive datasheets, the sequence and interlock specification, alarm and event logs, and a simulator or test harness for the process. - Tools: an IEC 61131-3 runtime or compiler such as OpenPLC or MATIEC, the process simulator, Python, PDF tools. - Put in everything a controls engineer 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 control logic 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 sequence, interlock or recipe fault on that cell. - The model works through the program and logs, changes the logic, runs it against the simulator, and writes the patched program plus answer.json naming the change. - Your marking script runs the patched program through the normal sequence and a set of hidden fault scenarios, such as a late sensor, a drive trip, power loss mid-cycle or an operator restart. The score is the share of scenarios that end in a safe and correct state. - A strong task lets the naive fix pass the normal sequence and fail a fault path. - 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 - Sequence stalls when one sensor responds late. Interlock can be bypassed on restart. Recipe change breaks a dosing step. Next steps:Press "Apply"We will review your applicationIf qualified, you will be accepted into the network and can be considered for this and similar positions & projects We can only accept candidates that are willing to work as individual contractors / freelancers.