Benefits:
Core Applied AI Engineer – Agentic AI & Security (W2 POSITION ONLY, NO C2C OR 1099)
Introduction Our client is seeking a Core Applied AI Engineer to architect and deploy end-to-end agentic workflows that automate critical business processes. The ideal candidate pairs deep expertise in core applied AI - including reasoning, orchestration, context engineering, and harness engineering - with rigorous security engineering. This role moves far beyond traditional RAG or conversational chatbots. You will build secure, autonomous agents capable of reasoning, decision-making, API/tool invocation, and safe integration within complex enterprise environments.
Responsibilities
Agentic Architecture & Orchestration
- End-to-End Workflow Automation: Design and deploy AI agents that execute complex business processes across enterprise APIs, databases, SaaS platforms, and MCP/tool servers.
- Reliability & Optimization: Utilize context engineering and harness engineering to optimize agent performance, control, observability, and traceability.
- Cross-Functional Collaboration: Partner with Product, Data, Platform, and Security teams to embed AI capabilities seamlessly into enterprise infrastructure.
Agentic Security & Governance
- Security Controls & Guardrails: Design robust guardrails governing agent tool invocation, API access, data retrieval, and autonomous decision-making.
- Threat Mitigation: Protect systems against agent-specific vulnerabilities, including direct/indirect prompt injection, context poisoning, tool abuse, privilege escalation, and data exfiltration.
- Identity & Access Management: Implement least-privilege access models, service identities, OAuth/OIDC, RBAC/ABAC, and secure secrets management.
- Human-in-the-Loop: Establish automated and manual checkpoints requiring approval where business or security risk thresholds demand it.
Observability, Testing & Evaluation
- Telemetry & Auditing: Implement tracing and monitoring frameworks to log agent reasoning, decision paths, tool calls, and downstream actions.
- Adversarial Testing: Establish rigorous evaluation pipelines to test agent behavior, security boundaries, and failure modes under adversarial scenarios.
Requirements
Required Skills
- Engineering Experience: Proven track record of building and deploying production-grade agentic AI systems or autonomous workflows.
- Core AI Expertise: Deep proficiency in context engineering, harness engineering, reasoning loops, and agent orchestration.
- Security Background: Strong foundation in application, API, data, or cloud security, with specific experience securing AI architectures.
- Enterprise Security Knowledge: Familiarity with LLM threat vectors (prompt injection, excessive agency, data leakage) alongside traditional IAM, least-privilege, and policy enforcement models.
- Analytical & Communication Skills: Excellent problem-solving capabilities with the ability to bridge technical and business requirements effectively.
Preferred Skills
- Experience with Model Context Protocol (MCP), multi-agent architectures, or agentic frameworks.
- Background in AI security frameworks, threat modeling, red teaming, or adversarial testing.
- Proficiency in cloud security (Azure, AWS, or Google Cloud Platform) and large enterprise data platforms.
- Experience integrating AI agents with transactional systems such as ERP, CRM, or commerce platforms.
Key Interview Discussion Areas
Candidates for this role should be prepared to discuss hands-on experience addressing:
- Determining and enforcing strict operational boundaries and permissions for an AI agent.
- Securing an agent's context, memory, tool servers, and execution environment.
- Detecting and mitigating prompt injection and context manipulation.
- Designing audit trails and telemetry for agent reasoning and tool execution.
- Constructing adversarial testing frameworks to validate agent reliability.
This is a remote position.