Member of Technical Staff, ML Security

present of AI; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Higgsfield is seeking a security-focused infrastructure engineer to own security for GPU clusters and ML systems that train and serve the company's models. You will harden Linux hosts, Kubernetes, and container setups, and ensure strong isolation between tenants and workloads while protecting model weights and customer data.

You will work with outside providers to enforce clear security standards, perform risk assessments, and respond to incidents as part of building a robust AI platform

Qualifications

• Deep infrastructure security experience.
• Hands-on time with GPU or distributed training setups.
• Ability to hold outside providers to clear security standards.

Responsibilities

• Own security for GPU clusters and ML systems that train and serve the company's models.
• Hardening Linux hosts, Kubernetes and container setups, keeping tenants and workloads separated, and protecting model weights and customer data.
• Coordinate security assessments and incident response with internal teams and external providers.

Skills

Kubernetes hardening

Linux hardening

GPU security

Vendor security

IR & detection

Python/Go/Bash

Job description

This role owns security for the GPU clusters and machine learning systems that train and serve the company's models. That means hardening Linux hosts, Kubernetes and container setups, keeping tenants and workloads separated, and protecting model weights and customer data. A strong candidate brings deep infrastructure security experience, hands-on time with GPU or distributed training setups, and the ability to hold outside providers to clear security standards.

How it fits the AI era

AI workloads bring new things to protect, such as model weights, training data, and shared GPU hardware, and new threats like model theft or unauthorized GPU use. Tooling can help with scanning, detection, and scripting routine checks. Judgment on risk, vendor trust, and balancing safety against speed and reliability stays with people.

How it builds AI experience

Someone in this role would get close exposure to how a large video and image generation product is trained and served, from GPU clusters to live inference. Building toward it means practicing with Kubernetes GPU setups, learning how training and inference differ, and studying how model files and datasets get moved and stored. Reviewing the security of outside GPU providers adds a view into how the wider AI supply chain works.

Worth showing

• Kubernetes and Linux hardening
• GPU cluster security work
• Vendor security assessments
• Detection and incident response
• Python, Go, or Bash scripting

This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.

About Higgsfield

Higgsfield is an AI-native creative suite that generates cinematic-quality video and images from text, image, or link inputs, per the company.