Senior AI Security Engineer - Agentic Systems

synapxe; Singapore; On-site; SGD 120,000 - 180,000; Full time

Job summary

synapxe in Singapore is seeking an experienced AI engineer to design and build agentic AI harnesses for cybersecurity, defining the control plane, evaluation framework, memory architecture and guardrails to deploy safe AI across cyber functions.

The role covers integration with security tooling, incident response, threat intel, detection engineering, cloud security and IAM, with emphasis on secure tooling, auditability and production observability.

Qualifications

• Strong hands-on experience building production-grade LLM, agentic AI, ML, automation or platform systems.
• Deep understanding of agent architecture, orchestration frameworks, tool calling, memory design, RAG, model routing and multi-agent workflows.
• Experience with frontier models, open-source models or both, including evaluation, benchmarking and model comparison.
• Strong software engineering background, including Python, APIs, backend services, cloud platforms, containers, CI/CD, authentication, logging and production observability.
• Experience integrating AI systems with enterprise APIs, identity systems, data platforms, workflow engines, ticketing systems, code repositories and operational tools.
• Prior experience operating or supporting production systems, including monitoring, alerting, incident response, rollback, release management, access control, cost management and post-incident review.
• Practical understanding of production failure modes such as model drift, prompt regressions, broken tool calls, API failures, retrieval errors, permission issues, latency problems, data quality gaps, cost spikes and unsafe outputs.
• Practical understanding of AI safety risks, including hallucination, prompt injection, insecure tool use, excessive agency, sensitive data leakage, memory poisoning, adversarial manipulation and unsafe autonomous behaviour.
• Experience designing human-in-the-loop workflows for high-risk, regulated or security-sensitive environments.
• Ability to design for operational handover, including runbooks, support models, service ownership, observability, change control and measurable service health.

Responsibilities

• Design secure agentic AI architectures supporting planning, reasoning, tool use, memory, retrieval, model routing, multi-agent coordination and human-in-the-loop workflows.
• Build the agent harness and control plane to manage autonomy, policies, approvals, audit logs, rollback, kill switches and agent action boundaries.
• Implement AI-to-cyber integrations with SIEM, SOAR, EDR, IAM, PAM, CMDB, ITSM, scanners, cloud platforms, repositories, CI/CD, ticketing and knowledge systems.
• Build secure tool mediation to define what agents can read, recommend, draft, test, execute or elevate, with approvals for high-risk actions.
• Define agent identity and access controls using least privilege, scoped credentials, JIT access, secrets isolation, session boundaries and full auditability.
• Secure the agentic AI supply chain across prompts, tools, connectors, MCP servers, plugins, packages, containers, models, datasets and retrieval sources.
• Build the cyber data, memory and knowledge layer using RAG, vector search, knowledge graphs, case memory and context stores.
• Ensure agent outputs are evidence-based, traceable to source systems, alerts, logs, tickets, vulnerabilities, threat intelligence and case notes.
• Develop reusable cyber agent patterns for triage, investigation, threat intelligence, vulnerability analysis, secure code review, detection, GRC and remediation.
• Evaluate frontier and open-source models for reasoning, coding, tool use, cyber performance, reliability, hallucination, latency, cost, safety and deployment fit.
• Design model-agnostic architecture to support model routing, fallback, regression testing, cost controls, latency targets and graceful degradation.
• Build AI evaluation and test harnesses covering benchmarks, adversarial tests, simulations, incident replay, human review and operational acceptance criteria.
• Create cyber simulation environments to safely test agents against historical incidents, SOC cases, vulnerable code, phishing, cloud attack paths and red-team scenarios.
• Design controls against prompt injection, malicious documents, poisoned tickets, hostile webpages, compromised retrieval sources and memory poisoning.
• Build authorised AI-assisted cyber assessment capabilities for code review, vulnerability discovery, exploit validation, patch suggestions, testing and red-team planning.
• Define human decision rights, ownership, approvals, monitoring responsibilities and clear boundaries where AI can assist versus where humans must decide.
• Design for production operations, including monitoring, logging, rollback, runbooks, ownership, access reviews, cost controls, LLMOps and lifecycle management.
• Partner with cyber SMEs to convert operational workflows into safe, measurable, production-grade AI capabilities with clear controls and escalation paths.

Skills

Production-grade LLM

Agentic AI

ML

Automation

Platform systems

Agent architecture

Orchestration frameworks

Tool calling

Memory design

RAG

Model routing

Multi-agent workflows

Frontier models

Open-source models

Evaluation

Benchmarking

Model comparison

Python

APIs

Backend services

Cloud platforms

Containers

CI/CD

Authentication

Logging

Production observability

Enterprise APIs

Identity systems

Data platforms

Workflow engines

Ticketing systems

Code repositories

Operational tools

Monitoring

Alerting

Incident response

Rollback

Release management

Access control

Cost management

Post-incident review

AI safety risks

Hallucination

Prompt injection

Insecure tool use

Excessive agency

Data leakage

Memory poisoning

Adversarial manipulation

Unsafe autonomous behaviour

Human-in-the-loop

Runbooks

Service ownership

Observability

Change control

Measurable service health

Tools

Python

APIs

Backend services

Cloud platforms

Containers

CI/CD

Authentication

Logging

Production observability

Job description

synapxe in Singapore is seeking an experienced AI engineer to design and build agentic AI harnesses for cybersecurity, defining the control plane, evaluation framework, memory architecture and guardrails to deploy safe AI across cyber functions.

The role covers integration with security tooling, incident response, threat intel, detection engineering, cloud security and IAM, with emphasis on secure tooling, auditability and production observability.