SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 276,960 - 363,510 / annual
Suno is building the world's first creative entertainment platform for music creation, backed by leading investors including NVIDIA Ventures, Union Square Ventures, and Lightspeed Venture Partners. As a Staff AI Security Engineer, you will secure Suno's AI models across their entire lifecycle—from training pipelines and compute infrastructure to the products used by millions of creators.
You'll design and build security controls for model training and serving infrastructure, including compute clusters, pipelines, access to model artifacts, and the software supply chain. You'll engineer detection, monitoring, and guardrail systems for model inputs, outputs, and agentic tool use as production code. Your work will include adversarial testing and red-teaming across models, pipelines, and products, then driving fixes to completion.
You'll partner closely with ML research, platform, and product engineering teams to make secure defaults the easy path in model training, evaluation, and deployment. You'll threat model and review customer-facing AI features before launch, and build shared controls that enable teams to ship safely by default. You'll set technical direction for AI security at Suno, including standards, reference architectures, and roadmap, while raising the security bar across engineering.
This is a high-impact role at a fast-growing company where you'll shape a new medium and work with a small team that cares deeply about quality.
REQUIREMENTS:
- 8–10 years in security engineering, including hands-on work securing ML systems or large-scale ML infrastructure in production
- Track record of designing and shipping security systems and tooling, not only assessing them
- Deep knowledge of cloud and infrastructure security: IAM, Kubernetes, network segmentation, secrets management, and infrastructure as code
- Strong understanding of ML training and serving pipelines and their attack surface, including model supply chain risks and protecting model weights and artifacts
- Fluency with LLM and generative AI attack vectors and frameworks such as OWASP Top 10 for LLMs and MITRE ATLAS
- Strong coding skills in Python or Go
- Experience setting technical direction and influencing senior engineers and leaders without formal authority
NICE-TO-HAVES:
- Experience at a foundational model company or an AI research lab
- Securing GPU compute environments or distributed training at scale
- Hands-on use of adversarial testing tools such as Garak or PyRIT
- Background securing audio, media, or other generative content pipelines
