Principal AI Engineer

Role Overview

• Hands-on Principal AI Engineer focused on building and productionising agentic AI systems for cybersecurity.
• Core focus is AI engineering, not cybersecurity expertise.
• Build production-grade LLM, agentic AI, ML, automation and platform systems.
• Cyber domain expertise will be provided by security SMEs.
• Must have experience designing/operating reliable, observable, auditable, recoverable and supportable production systems.

Core Technical Scope

• Agentic AI architecture: planning, reasoning, tool use, task decomposition, memory, RAG, model routing and multi-agent workflows.
• Agentic AI harness & control plane:Autonomy levelsPolicy enforcementApproval workflowsAudit loggingRollbackKill switchesAction limits
• AI-to-system integration with enterprise/cyber tools such as:SIEM / SOAREDR / NDRIAM / PAMCMDB / ITSMVulnerability scannersCloud security platformsCode repositories / CI/CDTicketing and knowledge platforms
• APIs, connectors, webhooks, queues, MCP, service accounts, scoped credentials, session controls and rate limiting.
• Secure tool mediation and clear boundaries between read / recommend / draft / execute / escalate.

AI Security & Governance

• Agent identity, authentication, authorisation and least-privilege access.
• JIT access, scoped credentials, secrets isolation and approval-bound permissions.
• Secure AI supply chain covering:PromptsTools/connectorsMCP serversPlugins/skillsPackages/containersModelsEvaluation datasetsRetrieval sources
• Provenance, allowlisting, signing, dependency scanning, sandboxing and change control.
• Controls against:Prompt injectionIndirect prompt injectionData leakageMemory poisoningTool manipulationMalicious documents/web contentExcessive agencyUnsafe autonomous behaviour

Data, Memory & Knowledge

• Build RAG, vector search, knowledge graphs, structured knowledge and case memory.
• Work with cyber data such as:Assets and identitiesVulnerabilitiesAlerts/incidentsThreat intelligenceControls/playbooksTicketsSource code/dependenciesInvestigation history
• Implement evidence provenance, source trust, freshness and confidence controls.

Agent Development

• Build reusable agents/patterns for:Alert triageIncident investigationThreat intelligenceVulnerability analysisSecure code reviewDetection engineeringGRC/control testingRed-team supportRemediationExecutive reporting
• Move AI from POCs/demos → reusable, measurable production capabilities.

Models & Evaluation

• Evaluate frontier and open-source LLMs.
• Benchmark:ReasoningCodingTool useCyber-task performanceHallucinationReliabilityLatencyCostContext handlingSafety
• Design model-agnostic architectures supporting model routing, fallback and graceful degradation.
• Build AI evaluation/test frameworks covering:AccuracyFalse positives/negativesHallucinationPrompt injectionData leakageUnsafe tool useMemory poisoningFailure recoveryReliability

Simulation & Testing

• Build controlled environments for:Historical incident replaySynthetic SOC casesVulnerable codeCloud attack pathsPhishing scenariosDetection engineeringGRC workflowsRed-team simulations
• Support authorised AI-assisted security testing such as code review, vulnerability discovery, exploitability validation and attack-path analysis.

Production & LLMOps

• Design for:MonitoringLoggingAlertingRollbackRunbooksService ownershipAccess reviewsCost controlsOperational handover
• Implement LLMOps / agent lifecycle management.
• Manage prompts, agents, tools, model versions, evaluations, telemetry, drift, releases and continuous improvement.
• Understand production failure modes: model drift, prompt regression, broken tool calls, API failures, retrieval issues, permission failures, latency, data quality and cost spikes.

Must-Have Experience

• Strong hands-on production-grade LLM / Agentic AI / ML / automation / platform engineering.
• Strong knowledge of:Agent architectureOrchestrationTool callingMemoryRAGModel routingMulti-agent systems
• Experience with frontier and/or open-source models and model evaluation.
• Strong software engineering:PythonAPIsBackend servicesCloudContainersCI/CDAuthenticationLoggingObservability
• Enterprise API and system integration experience.
• Prior production operations/support experience.
• Understanding of AI security risks and human-in-the-loop controls.
• Ability to design systems for operational handover and support.

Disclaimer: The company is committed to ensuring theprivacy and security of your information. By submitting this form, you consentto the collection, processing, and retention of the information you provide.The data collected (which may include your contact details, educationalbackground, work experience and skills) will be used solely for the purpose ofevaluating your qualifications for the position you're applying for. Your datawill be stored securely and retained for the duration necessary to fulfil ourhiring process. If you are not selected for the position, your data will bekept on file for a limited period in case future opportunities arise. You havethe right to access, correct, or delete your data at any time by contacting usat Quess Singapore | A Leading Staffing Services Provider in Singapore(quesscorp.sg)

“This is in partnership with the Employment andEmployability Institute Pte Ltd (“e2i”).

e2i is the empowering network for workers andemployers seeking employment and employability solutions. e2i serves as abridge between workers and employers, connecting with workers to offer jobsecurity through job-matching, career guidance and skills upgrading services,and partnering employers to address their manpower needs through recruitment,training, and job redesign solutions. e2i is a tripartite initiative of theNational Trades Union Congress set up to support nation-wide manpower andskills upgrading initiatives. By applying for this role, you consentto Quesscorp Singapore’s PDPA and e2i’s PDPA.”