Usage Scenarios
Engee MIND помогает выполнять инженерные задачи через агентный клиент. Вы задаете требуемый результат, предоставляете исходные данные и определяете допустимые изменения. Агент выбирает навыки, выполняет действия в Engee and returns verifiable artifacts. Below is a small but exhaustive list of engineering scenarios that can be implemented with Engee MIND.
| Scenario | What to prepare | Expected result |
|---|---|---|
Description of the system or existing model, parameters and criteria |
Model |
|
Component documentation, measurements if available |
Script and block models of acceptable levels, graphs and report |
|
Simulink conversion and verification |
|
Model Engee and the comparison protocol |
Application development Engee |
User scripts, calculated functions, project or repository |
Genie App with interface, calculation logic and checks |
The computer for the integration client and the available local resources |
Connecting to the current environment Engee with verification results |
|
Machine name/IP, model, mode and test conditions |
Configured communication, deployed model, logs and test data |
Task statement template
Цель: [какой инженерный результат нужен].
Исходные данные: [файлы, путь проекта, документация, параметры и единицы].
Разрешенные изменения: [что можно создавать или менять].
Ограничения: [что сохранять, чего не менять, допустимые средства].
Проверка: [эталон, диапазон, метрики и допуски].
Результат: [модель, код, данные, графики, отчет и место сохранения].
The general order of work
-
Check the connection Engee MIND and access to the user environment.
-
Transfer the input files. An attachment to a chat does not guarantee that the file has already been uploaded to the session Engee.
-
Ask the agent to clarify only those missing data that affect the correctness of the result.
-
Get an action plan and a list of skills used.
-
Wait for the actual calculations or tests and check the criteria.
-
Get the paths of the saved files and the necessary files via
file_download.
What to check as a result
It should be clear which source data has been used, what has been changed, which calculation has been performed and which criteria have been passed. Source data limitations, unsupported blocks, failed tests, and approximate assumptions should be listed explicitly.
For repeatable procedures, you can create custom skill, so that the next task is performed according to the same methodology.