DeepSeek is hiring 150 engineers, and none of them will touch a model
Hundreds of thousands of AI agent sandboxes can already run concurrently on a single DeepSeek cluster. Now the company is The post DeepSeek is hiring 150 engineers, and none of them will touch a model appeared first on The New Stack .
DeepSeek is expanding its engineering team by hiring 150 engineers on September 7, with a focus on server-side engineering and Agent Elastic Compute. These engineers will be responsible for upgrading, maintaining, and rewriting DeepSeek's existing backend systems as workloads grow. The company's Elastic Compute system, or DSec, is a sandbox infrastructure used to execute agent workloads during post-training and evaluation.
DSec supports four types of environments for agent workloads using Python SDK, ranging from simple function calls to full operating system environments. To manage storage and network bandwidth consumption, DeepSeek ties DSec into its 3FS distributed filesystem and keeps container base images and filesystem commits as read-only layers backed by 3FS.
This allows for efficient lazy loading of agent environments and reduces duplicate page-cache usage across virtualized environments, reclaiming memory to allow safe overcommitment. Additionally, DSec uses a globally ordered trajectory log to record commands and their results, enabling fast-forwarding and replaying of interrupted training and evaluation runs.
Written by urgent.news from The New Stack's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
