AI Engineer - Cybersecurity Products

About the role

Assurity Trusted Solutions (ATS), a wholly owned subsidiary of GovTech, is hiring an AI Engineer to build agentic AI systems for cybersecurity use cases. The role blends large language models with solid AI/ML fundamentals, including data pipelines, classical machine learning, rigorous evaluation and safety guardrails, to deliver reliable, auditable services. It is a direct contract until 31 March 2027, subject to extension based on performance.

Key responsibilities

• Design, build and ship agentic AI features (planning and execution loops, tool use and function calling, multi-step workflows) for use cases such as vulnerability triage, exploit reproduction assistance and incident-response copilots.
• Implement and harden retrieval-augmented generation (RAG): indexing, chunking, routing, re-ranking, feedback loops and data governance for sensitive environments.
• Set up evaluation and observability for LLM and agent workflows (tracing, cost/latency/quality dashboards, offline and online evaluations, guardrail hit rates) and turn insights into product changes.
• Build safety and guardrails (content policies, schema and output validation, PII redaction, prompt-injection and jailbreak defenses, tool permissioning) and monitor them in production.
• Apply traditional machine learning (classification, regression, anomaly detection) where it is simpler or more effective, running A/B tests and error analysis.
• Own productionization: CI/CD for AI apps, containerization, scalable inference endpoints, vector and search infrastructure, runbooks and SLOs.
• Collaborate with product and security teams to scope problems, write design documents and meet security and privacy requirements.

Required profile

• 2 or more years in AI/ML/LLM engineering or software engineering with meaningful AI/ML contributions; exceptional internships, research or startup projects are welcome.
• Proven delivery of at least one production AI feature, ideally an LLM/agent workflow with RAG, evaluation and monitoring.
• Clear communicator able to scope problems, write concise design proposals and work with product managers and security stakeholders.
• Working knowledge of cybersecurity concepts (e.g. OWASP Top 10, OAuth/OIDC, data protection) and motivation to solve security problems with AI; newcomers to cybersecurity are welcome.

Required skills

• LLM and agent systems with frameworks such as LangChain, AutoGen or CrewAI, and RAG stacks (embedding models, vector stores, hybrid search, re-rankers, prompt templating, caching).
• Evaluation harnesses (golden sets, rubric/LLM judges), tracing (OpenTelemetry, Langfuse) and dashboards; A/B testing.
• Safety and guardrails: policy enforcement, output validation, least-privilege tool access, prompt-injection mitigation, secrets handling and PII redaction.
• Machine learning fundamentals with sklearn, XGBoost or LightGBM, including cross-validation and metrics.
• Cloud (AWS, Azure or GCP), Docker, Kubernetes, CI/CD, Terraform or CloudFormation, and SQL/NoSQL.
• Strong Python, unit and integration testing, code reviews, API design and documentation.

What we offer

• Remuneration commensurate with qualifications and experience.
• A yearly training budget and an annual performance bonus.
• Contract staff enjoy the same benefits as permanent employees.