AI Orchestration Engineer, Global Information Security (Vice President)

Company

Jefferies, the global investment banking firm, has served companies and investors for almost 60 years. Headquartered in New York with its European head office in London, the firm provides clients with capital markets and financial advisory services, institutional brokerage and securities research, and asset management. Jefferies provides research and execution services in equity, fixed income, foreign exchange, and a full range of investment banking services including underwriting, merger & acquisition, restructuring and recapitalisation and other advisory services, with businesses operating in the Americas, Europe and Asia.

Overview

Global Information Security is building an internal AI automation capability that transforms security processes into scalable production services. Our focus is on applying AI, workflow orchestration and software engineering to improve the effectiveness, consistency and efficiency of security operations and adjacent cyber domains.

This is a senior hands-on engineering role responsible for designing, building and operating AI-enabled automation used daily by analysts and engineers across multiple regions. The successful candidate will work primarily with the Security Operations Centre while supporting broader security functions including identity and access management, privileged access management, cloud security, network security, data protection and vulnerability management.

We are seeking an experienced engineer who can turn operational problems into production-ready solutions. Recent experience should include development of agentic systems, workflow orchestration, Model Context Protocol (MCP) integrations, APIs and automation platforms that deliver measurable outcomes in enterprise environments.

The role is based in London and aligned to support collaboration across both APAC and Americas teams.

Responsibilities

• Design, build and operate AI-enabled security automation services, treating them as production systems with appropriate testing, monitoring, alerting, resilience and lifecycle management.
• Deliver automation supporting security operations, including alert triage, enrichment, case summarisation, evidence collection, threat intelligence analysis, detection engineering support and repetitive investigative workflows.
• Develop and maintain MCP servers, APIs and integration services across SIEM, EDR, SOAR, PAM, IGA, CSPM, DLP, ticketing and threat intelligence platforms.
• Build agentic and workflow-driven solutions, making informed decisions on when AI is appropriate and when deterministic automation provides a better outcome.
• Design multi-step orchestration patterns including context management, tool selection, task decomposition and human approval workflows.
• Establish reusable integration patterns, shared services, prompt libraries, tool catalogs and evaluation frameworks that can be leveraged across multiple security domains.
• Partner with Cyber Operations leadership and domain leads to identify high-value automation opportunities and prioritise delivery against business outcomes.
• Engineer security and governance controls into all solutions, including least-privilege access, human-in-the-loop approvals, audit logging, traceability and resilience against prompt injection and tool misuse.
• Measure and report the effectiveness of delivered solutions through analyst time savings, cycle-time reduction, accuracy improvements and quality metrics.
• Mentor engineers within the global delivery organisation and contribute to a follow-the-sun operating model through documentation, standards and knowledge sharing.

Required qualifications

• 5 + years professional software engineering with production ownership
• Proficiency in Python or other programming languages, including asynchronous programming, API integration, automated testing and CI/CD.
• Demonstrated production experience building LLM-backed or agentic systems: retrieval, tool and function calling & orchestration.
• Recent, demonstrable delivery of agentic or LLM-backed systems in production: tool and function calling, orchestration of multi-step work, retrieval, and evaluation or regression testing of non-deterministic output.
• Experience integrating enterprise platforms through REST APIs, webhooks, event-driven patterns and authentication frameworks.
• Practical experience with at least one agentic development environment or agent runtime (for example Claude Code, Bedrock AgentCore, or equivalent) and one workflow automation platform (for example n8n, or equivalent commercial iPaaS).
• Cloud engineering experience (AWS or Azure), containerisation and infrastructure-as-code.
• Secure development practice: secrets management, least-privilege service identity, input validation, output encoding.
• Demonstrated delivery across multiple time zones for distributed stakeholders.

Preferred qualifications

• Financial services or comparable regulated industry experience, with working knowledge of audit, evidence and change-control requirements.
• Familiarity with agentic AI failure modes including prompt injection, tool poisoning, excessive agency, and current mitigations.
• Production observability practice: structured logging, distributed tracing, metrics and alerting and specifically experience instrumenting and evaluating non-deterministic systems, including LLM tracing or evaluation tooling (for example LangSmith, Langfuse, Phoenix, OpenTelemetry-based tracing, or equivalent).
• Practical familiarity with one or more further security domains: identity and access management, privileged access, network or cloud security, data protection, vulnerability management.
• Experience building evaluation and quality-monitoring pipelines for LLM systems, including judge-model calibration against human review.
• Experience transitioning vendor-managed automation to internally owned capability.
• Open-source contribution published research or conference speaking.

Candidate Profile

The successful candidate is a builder who can take a loosely defined problem and deliver reliable, supportable capability without requiring detailed direction. They are equally comfortable working with analysts, engineers and senior security leaders to understand operational challenges and design pragmatic solutions.

They understand that not every problem requires AI. They know when deterministic automation is the correct answer, when agentic workflows add value, and how to balance innovation with governance, security and operational reliability.

Above all, they measure success by business outcomes, operational adoption and long-term sustainability rather than demonstrations or prototypes.

#LI-JD1