- Популярные видео
- Авто
- Видео-блоги
- ДТП, аварии
- Для маленьких
- Еда, напитки
- Животные
- Закон и право
- Знаменитости
- Игры
- Искусство
- Комедии
- Красота, мода
- Кулинария, рецепты
- Люди
- Мото
- Музыка
- Мультфильмы
- Наука, технологии
- Новости
- Образование
- Политика
- Праздники
- Приколы
- Природа
- Происшествия
- Путешествия
- Развлечения
- Ржач
- Семья
- Сериалы
- Спорт
- Стиль жизни
- ТВ передачи
- Танцы
- Технологии
- Товары
- Ужасы
- Фильмы
- Шоу-бизнес
- Юмор
Enterprise AI Infrastructure Security - 4) Service Accounts & Automation Security
Securing ClearML for the Enterprise — Part 4: Service Accounts & Automation Security
In this video we walk through ClearML's service accounts — the identities behind your automated workloads — and how impersonation ensures least-privilege execution across your agents, pipelines, and schedulers.
What we cover:
- Why automated workloads need identities — agents, CI/CD pipelines, schedulers
- What a service account is and how it differs from a human user
- Creating a service account and assigning it to a group
- How group membership gives service accounts the same governance as users — access rules, administrator vaults, all inherited automatically
- The "one account per purpose" principle — scoping credentials to specific roles
- Impersonation with useOwnerToken: true — how agents execute tasks as the submitting user, not as themselves
- Walking through real agent logs showing impersonation in action — owner tokens and per-user vault loading
- When to impersonate vs when to let a service account run as itself
- Credential best practices — secrets managers, rotation schedules, deleting unused accounts
- How ClearML redacts credentials in logs by default
Previous videos in this series:
- Part 1 — Introduction to the Six Layers of Enterprise Security: https://www.youtube.com/watch?v=RXz8FuzzwI4
- Part 2 — Identity Provider Setup, Group Sync & Access Rules: https://www.youtube.com/watch?v=wnVbOxWbzWM
- Part 3 — Configuration Governance with Administrator Vaults: https://youtu.be/vse_015TaWM
This is Part 4 of our series on enterprise AI infrastructure security. Whether you're an IT director evaluating ClearML, a platform engineer rolling it out, or a security architect designing your automation layer — this walkthrough covers the practical, hands-on configuration from start to finish. We have also written a blog about this video here: https://clear.ml/blog/service-accounts-and-automation-security-with-clearml
🔗Links & Resources
ClearML Enterprise: https://clear.ml/enterprise ClearML Docs — Service Accounts: https://clear.ml/docs/latest/docs/user_management/service_accounts/ ClearML Docs — Administrator Vaults: https://clear.ml/docs/latest/docs/user_management/admin_vaults/ ClearML Docs — User Groups: https://clear.ml/docs/latest/docs/user_management/user_groups/ ClearML Docs — K8s Glue Agent Configuration: https://clear.ml/docs/latest/docs/clearml_agent/clearml_agent_k8s_glue/
Видео Enterprise AI Infrastructure Security - 4) Service Accounts & Automation Security канала ClearML
In this video we walk through ClearML's service accounts — the identities behind your automated workloads — and how impersonation ensures least-privilege execution across your agents, pipelines, and schedulers.
What we cover:
- Why automated workloads need identities — agents, CI/CD pipelines, schedulers
- What a service account is and how it differs from a human user
- Creating a service account and assigning it to a group
- How group membership gives service accounts the same governance as users — access rules, administrator vaults, all inherited automatically
- The "one account per purpose" principle — scoping credentials to specific roles
- Impersonation with useOwnerToken: true — how agents execute tasks as the submitting user, not as themselves
- Walking through real agent logs showing impersonation in action — owner tokens and per-user vault loading
- When to impersonate vs when to let a service account run as itself
- Credential best practices — secrets managers, rotation schedules, deleting unused accounts
- How ClearML redacts credentials in logs by default
Previous videos in this series:
- Part 1 — Introduction to the Six Layers of Enterprise Security: https://www.youtube.com/watch?v=RXz8FuzzwI4
- Part 2 — Identity Provider Setup, Group Sync & Access Rules: https://www.youtube.com/watch?v=wnVbOxWbzWM
- Part 3 — Configuration Governance with Administrator Vaults: https://youtu.be/vse_015TaWM
This is Part 4 of our series on enterprise AI infrastructure security. Whether you're an IT director evaluating ClearML, a platform engineer rolling it out, or a security architect designing your automation layer — this walkthrough covers the practical, hands-on configuration from start to finish. We have also written a blog about this video here: https://clear.ml/blog/service-accounts-and-automation-security-with-clearml
🔗Links & Resources
ClearML Enterprise: https://clear.ml/enterprise ClearML Docs — Service Accounts: https://clear.ml/docs/latest/docs/user_management/service_accounts/ ClearML Docs — Administrator Vaults: https://clear.ml/docs/latest/docs/user_management/admin_vaults/ ClearML Docs — User Groups: https://clear.ml/docs/latest/docs/user_management/user_groups/ ClearML Docs — K8s Glue Agent Configuration: https://clear.ml/docs/latest/docs/clearml_agent/clearml_agent_k8s_glue/
Видео Enterprise AI Infrastructure Security - 4) Service Accounts & Automation Security канала ClearML
Комментарии отсутствуют
Информация о видео
18 марта 2026 г. 20:48:55
00:11:07
Другие видео канала


![[OLD] Quick Introduction](https://i.ytimg.com/vi/-9vqxF2UfFU/default.jpg)

![[MLOps] The Clear SHOW - S02E11 - DIY Strikes Back! Building the Model Store!](https://i.ytimg.com/vi/WIZ88SmT58M/default.jpg)


![[MLOps] The Clear SHOW - S02E10 - Everything You Wanted To Know About Model Stores*](https://i.ytimg.com/vi/xliX3IhNdmw/default.jpg)






![[MLOPS] From #GTC21: Workshop - Demonstrating an End-to-End Pipeline for ML/DL Leveraging GPUs](https://i.ytimg.com/vi/SLNXo7aIjIU/default.jpg)


![[MLOPS] From GTC22: A journey towards building a scalable AI serving solution](https://i.ytimg.com/vi/s3BxtzjzDkk/default.jpg)



