Job description
- This role does the hands-on architecture work
- You design the secure patterns engineering builds against, review hard problems in depth, and help make secure-by-design real across the Hyperblock and Polaris platform and the AI product surface
- You work closely with the Principal Security Architect, who sets the technical bar
- You own significant areas of that work and deliver them end to end
- Engineers adopt good architecture because it is the better path
- You will build that kind of architecture, working on problems that carry real risk
- Anaplan is not a simple SaaS application
- The platform performs highly dimensional, enterprise-scale modelling with responsive recalculation and strong correctness expectations
- The technical surface includes the legacy Hyperblock engine, the newer Polaris engine, and a growing set of AI products
- These include CoModeler for AI-assisted model building and CoPlanner for conversational planning support
- Enterprise customers in regulated industries increasingly ask a hard question: if Anaplan is breached, can Anaplan staff or an attacker read their data?
- Answering it takes rigorous architecture work
- We are commissioning an independent Trail of Bits assessment of exactly this
- It covers end-to-end tenant isolation, the proposed BYOK design, and internal infrastructure boundaries
- You will help turn findings like these into concrete design changes
- That includes moving key management from legacy HSM-based designs to Vault, and extending encryption coverage to storage and data paths that lack an agreed architecture today
- Produce reference architectures: design and maintain security patterns for platform and AI components, working to the standards set by the Principal
- Contribute to identity and authorisation design: including tenant isolation, service-to-service trust, and authorisation for AI agents and analyst-style workflows
- Apply cryptography and key management: work on encryption, BYOK and key custody. Be clear about where data is plaintext, who can use keys, and what each control protects against
- Design AI guardrails: input and output controls, tool and authorisation boundaries, and prompt-injection and abuse defences. Draw on MITRE ATLAS and the OWASP LLM Top 10
- Run threat models: lead them on high-risk systems, working with engineers to trace trust boundaries and compromise paths. Turn the findings into prioritised, actionable design changes
- Support external assessments: prepare architecture evidence, join discovery sessions with engineering, and check that what is built matches what is documented
- Improve the supply chain and attack surface: build integrity and provenance, hardened images, and less exposed surface
- Stay hands-on: write and review designs and prototypes, and test your own work
- Grow the practice: mentor engineers, contribute to standards, and help turn design decisions into enforceable guidance
- A working understanding of AI and ML security, and of how LLM-driven features change identity, authorisation, supply chain, and verification
- The ability to design and review real systems, and to defend your decisions to strong engineers
- A record of influencing engineers without formal authority, including when risk and delivery pressure conflict
- Solid experience in secure software architecture, with depth in at least two of: cryptography and key management, identity and authorisation, tenant isolation, and threat modelling
- Experience with enterprise-scale SaaS or cloud-first systems, ideally multi-tenant
- An offensive grounding that you use to pressure-test designs
- Contributions to NIST, ISO, or OWASP, or talks and published research
- Experience designing for regulated environments, such as FedRAMP
- Experience with Vault, KMS/HSM, or BYOK/HYOK designs
- Clear written and verbal communication. You can explain architecture honestly to engineers, leadership, and customers
Is there something wrong with this job listing? Let us know.