Senior Offensive Security Engineer (Red Team)

Job Title: Senior Offensive Security Engineer, Red Team

The Experience

The Product Red Team focuses on attacking Salesforce's full deployed product ecosystem and surrounding infrastructure — not the corporate environment. As a Senior Offensive Security Engineer, you will work closely with the Red Team Director to design and run adversary simulations against product trust boundaries, uncovering high-impact attack chains and validating how weaknesses can be exploited end to end. This role emphasizes multi-step exploitation, depth, and systemic impact, and you'll help translate red team operations into engineering and detection improvements across a large, interconnected software-as-a-service (SaaS) product portfolio.

What You'll Actually Be Doing

• Build and demonstrate multi-step attack chains across the product ecosystem, spanning web application and application programming interface (API) abuse, identity and authorization failures, cross-tenant isolation breaks, and integration-plane trust boundary abuse (AppExchange packages, partner APIs, OAuth applications, MuleSoft connectors).
• Own engagements end to end: charter development, threat modeling, ranked hypothesis backlogs, and technical reporting with root causes, variant analysis, and systemic remediation guidance.
• Collaborate with Detection Engineering, Detection and Response, and Incident Response teams to validate alerting and visibility on emulated chains, lead readout sessions, and run purple-team operations that close visibility gaps.
• Apply AI-driven testing techniques — agentic tooling, automated reconnaissance, AI-assisted chain development — alongside traditional manual tradecraft, and dedicate roughly 10-20% of time to research on novel attack techniques and tooling.

You're Our Person If...

• You have deep, hands-on offensive security experience with a demonstrated focus on web application and API attack surfaces, including a history of finding multi-step chains (prior product or SaaS red team work, SaaS bug bounty Hall of Fame recognition, public Common Vulnerabilities and Exposures (CVEs), or equivalent — single-bug submissions alone don't meet this bar).
• You have strong command of web application and API exploitation (auth bypasses, business logic abuse, server-side request forgery (SSRF), injection classes, insecure direct object references (IDOR), OAuth/Security Assertion Markup Language (SAML)/OpenID Connect (OIDC) attacks), multi-tenant isolation failures, and identity federation (SAML, OIDC, OAuth scope confusion, System for Cross-domain Identity Management (SCIM), JSON Web Token (JWT) manipulation).
• You can read production code at scale in at least one language (Java, Python, Go, JavaScript, TypeScript, or Apex) and are fluent in cloud security for at least one major provider (IAM abuse, instance metadata, cross-account pivots, container and orchestration boundaries).
• You're familiar with DevOps and build pipeline attack surfaces (continuous integration/continuous delivery (CI/CD) abuse, dependency confusion, supply chain compromise, infrastructure-as-code (IaC) misconfigurations), and write clearly enough to hold up for a vice president (VP)-level engineering audience without losing technical precision.

Even Better If...

• You hold offensive security certifications such as Offensive Security Web Expert (OSWE), GIAC Cloud Penetration Tester (GCPN), eLearnSecurity Web Application Penetration Tester eXtreme (eWPTXv2), Certified Web Exploitation Expert (CWEE), or equivalent.
• You specialize in AI/ML attack surfaces (prompt injection, agentic privilege abuse, tool and retrieval-augmented generation (RAG) pipeline attacks), supply chain and build-pipeline attacks, mobile, or low-level exploitation.
• You've worked against managed package ecosystems, SaaS integration platforms, or multi-tenant SaaS isolation boundaries, and are familiar with MITRE ATT&CK, MITRE ATLAS, and the OWASP Top 10 as engagement planning references.
• You have experience operating under formal engagement charters and rules of engagement, with judgment about when a chain has reached an authorization boundary and work should stop.