AI Engineer

My Financial Services client is seeking to recruit a AI Engineer on an initial 6 month contract based in London. It is hybrid and will require 3x days onsite per week.

As a Back-End AI Engineer, you will design and deploy secure, scalable AI services that power next-generation use cases across client intelligence, document processing, and risk management. You'll work in a greenfield environment, building compliant AI pipelines using Gemini (GCP), Azure OpenAI or Self Hosting embedding security and privacy controls from experimentation to production, in alignment with the bank's cybersecurity and regulatory standards.

Accountabilities & Responsibilities

Architect and implement secure AI services from lab to production, ensuring scalability and compliance
Develop robust APIs for LLMs, RAG pipelines, agentic workflows and document intelligence systems
Embed cybersecurity and data privacy controls across all AI workflows (e.g., encryption, anonymisation, access logging)
Collaborate with the CISO function on threat modeling, security reviews, and AI-specific control design.
Integrate with enterprise IAM systems, enforcing RBAC, least privilege
Conduct vulnerability scans, pen-test remediation, and support internal and regulatory audits (FCA, PRA)

Required Knowledge & Experience

Delivered greenfield AI systems in production with secure-by-design architecture
Designed and managed AI lab environments using IaC, containerisation, and secure networking practices
Hands-on experience with LLM implementation, including fine-tuning, prompt engineering, and secure deployment
Built agentic workflows using modular LLM agents with memory, planning, and tool integration
Implemented Model Context Protocol (MCP) to manage secure, auditable context injection across agentic systems
Experience building RAG pipelines with strict data governance and contextual integrity
Familiarity with EU AI Act, FCA cybersecurity principles, and oversight of critical systems
Worked directly with cybersecurity and compliance teams in regulated deployments
Implemented or maintained controls under ISO 27001, NIST, or SOC2 frameworks

Technical Skills & Technologies:

Languages & Frameworks
Python (FastAPI), LangChain, Google AI SDK, Azure Open AI SDK
Cloud & AI Platforms
GCP: Vertex AI, Gemini API, Cloud Run, GCS, IAM, Secret Manager, Audit Logs
Azure: Azure ML, Azure OpenAI, Key Vault, Azure Policy
Experience with Self Hosting
LLM
Fine-tuning and prompt engineering for LLMs (e.g., GPT, Gemini, Claude)
Secure deployment of LLMs via APIs with input/output filtering and logging
Integration of LLMs into RAG pipelines, document intelligence, and agentic workflows
Use of vector databases (e.g., FAISS, Pinecone, Chroma) for semantic search and retrieval
Implementation of grounding, context injection, and response validation mechanisms
Model Context Protocol (MCP)
Implement secure, policy-aligned Model Context Protocol (MCP) for managing contextual memory, grounding, and session control in LLM-based systems
Enforce context boundary policies, context versioning, and traceability to support auditability and prevent data leakage
Integrate MCP with enterprise IAM and data governance frameworks to ensure compliant context injection and revocation
Agentic Workflows
Design and orchestrate agentic AI workflows using modular, goal-driven agents with memory, planning, and tool-use capabilities
Implement secure agent execution environments with task decomposition, tool chaining, and feedback loops
Integrate agents with enterprise systems (e.g., document stores, APIs, risk engines) while enforcing contextual integrity, rate limiting, and audit logging
Apply agentic patterns to automate complex financial tasks such as client onboarding, document summarisation, and risk signal extraction
Security Tooling
Static code analysis (Bandit, SonarQube)
Secrets scanning, encryption (at rest/in-transit), token management
Identity integration (Google Identity, Azure Entra ID)
Data Security & Governance
RAG pipelines with data classification, masking, and DLP
GDPR and data residency compliance
MLOps & DevSecOps
GitHub Actions, CI/CD security testing, model drift detection, audit logging
Lab Environment Tooling
Infrastructure-as-Code (IaC): Terraform, Pulumi
Containerization & Orchestration: Docker, Kubernetes (GKE/AKS)
Networking & Isolation: VPCs, private endpoints, firewall rules, network policies
Data Sandboxing: Synthetic datasets, masking, DLP tooling
Monitoring & Observability: Prometheus, Grafana, Cloud Logging