The Role:Fireworks runs multi-tenant GPU inference for hundreds of open models in production, and the security problems that come with it are concrete: isolating tenants on shared accelerators, protecting customer data and model weights across a multi-cloud footprint, and keeping all of it auditable while the platform grows fast. As Member of Technical Staff, Cloud & Infrastructure Security, you own the cloud, Kubernetes, and GPU infrastructure layers of that stack, the software supply chain behind them, and the detection and response that keeps it honest. You will work across teams to strengthen our posture and to make secure-by-default the easy path for every engineer here.
Key Responsibilities:- Design and build security-focused software and platform capabilities to protect customer data, models, and services across our multi-cloud infrastructure, including encryption, identity and access management, secure API gateways, secure model execution, and sandboxing strategies.
- Design and implement tenant isolation across shared GPU and accelerator infrastructure, covering host, device, scheduler, and model artifact boundaries.
- Perform security reviews of cloud-native architectures, including Kubernetes clusters, multi-cloud workloads, and distributed data stores, and build integrated systems for continuous security monitoring, anomaly detection, and automated response.
- Embed security into CI/CD pipelines using a DevSecOps approach, implementing automated scanning, policy enforcement, and secure-by-default build and deployment workflows.
- Apply a build-over-buy philosophy by designing and developing in-house security tooling and automation where it provides better control, scalability, and integration than off-the-shelf solutions.
- Build and operate a comprehensive vulnerability management program, partnering with various teams to remediate risks across applications, containers, cloud infrastructure, and dependencies.
- Operate and continuously improve security operations, including detection engineering, alert triage, incident response, and continuous improvement through post-incident reviews.
- Participate in red/blue team exercises, tabletop simulations, and post-incident root cause analysis to strengthen security resilience.
- Own the compliance posture for the infrastructure and product layers, starting with SOC 2 and ISO 27001, by embedding controls into the platform rather than documenting them after the fact.
Minimum qualifications:- 5+ years of experience in software engineering or security engineering with a strong focus on security, infrastructure, or cloud-native systems.
- Proficient in Python and/or Go, designing production-grade systems, and comfortable automating in Linux environments and in managed CI/CD (e.g., GitHub Actions, Harness, CircleCI).
- Production depth in GCP or AWS, including network segregation, authentication, authorization, encryption, data protection, intrusion detection, and cloud-specific security benchmarks.
- Hands-on experience with Kubernetes, Docker, and containerized production environments; deep knowledge of Kubernetes internals and native security controls is a strong plus.
- Working knowledge of modern identity and access controls (SAML, OAuth, OIDC, SSO, RBAC/ABAC), least-privilege design, and workload identity.
Preferred qualifications:- Experience designing, operating, and securing large-scale Kubernetes platforms, including control plane security, node hardening, and multi-tenant isolation.
- Experience designing, operating, and securing large-scale multi-cloud platforms and zero-trust architectures across AWS, GCP, Azure, Oracle Cloud, and GPU-as-a-service providers.
- Proficiency with infrastructure-as-code using Terraform and Python, including experience building modular policy-as-code frameworks.
- Experience designing detection engineering pipelines across cloud audit logs, network telemetry, and application signals.
- Experience taking a compliance framework such as SOC 2 or ISO 27001 from control design through audit.
- Familiarity with container image vulnerability remediation, SBOM generation, and software supply chain security.
