Principal AI Engineer

As the Principal AI Engineer, this role is a senior technical leader driving the design and delivery of the AI platform that underpins the organization's multi-year AI Workforce
Transformation. The role owns the end-to-end AI architecture – from the AI Orchestration Layer to a unified LLM Gateway – ensuring the platform seamlessly embeds AI across key workflows.

The position requires deep expertise in modern AI/ML platforms and enterprise architecture, operating at a senior level as an individual contributor. It exists to elevate operational productivity through data-centric AI enablement, leveraging unique datasets to build proprietary AI solutions that augment human capabilities.

Requirements

Technical Competencies

AI Platform Ownership

Design, build and own the end-to-end AI platform – AI Orchestration Layer, Unified LLM Gateway, vector stores and MCP integrations – using containerised microservices on Kubernetes (EKS/AKS).

Model Validation & Risk Control

Oversee a rigorous model-validation process; ensure each model (especially in credit and fraud) undergoes independent validation, bias testing and stress-testing with sign-off from Risk before deployment.

Documentation & Auditability

Maintain thorough documentation and audit logs for AI models and workflows – assumptions, training-data lineage, version history and limitations – so every AI outcome is traceable (auditability by design).

Standards Compliance

Ensure conformance to ISO 42001 and the NIST AI Risk Management

Framework, translating standards into internal policies for explainability, monitoring and periodic risk assessment.

Human-in-Loop to AI-in-Loop Transition

Govern the Human-in-the-Loop to AI-in-the-Loop transition – define criteria (accuracy ≥95%, high user trust, zero compliance issues) and hold the authority to revert systems to human supervision.

Vendor Due Diligence

Conduct technical due diligence on third-party AI tools and cloud services (SOC 2, encryption, zero data retention) with Procurement and InfoSec.

Privacy by Design

Implement privacy-by-design – using the AI Gateway to scrub PII before data reaches external models, and labelling AI-generated outputs where applicable.

Orchestration & Tooling

Configure and optimise AI orchestration platforms (e.g. LangChain, ML pipelines), LLM gateways across providers (OpenAI, Anthropic,

HuggingFace), and vector databases (Pinecone, Weaviate, FAISS).

MLOps / DevSecOps

Implement MLOps / DevSecOps pipelines (Jenkins, GitLab CI/CD, GitHub Actions, Terraform/CloudFormation) with integrated SAST/DAST security scanning.

Monitoring & Observability

Set up monitoring and observability (Grafana, ELK, PagerDuty) for response times, throughput, error rates and token usage.

Education & Experience

A Bachelor’s degree in Computer Science, Software Engineering or related field (a Master’s degree in AI/ML or Data Science is an added advantage), with 8+ years in software engineering or architecture – including at least 3–5 years designing AI, data or cloud architectures at scale and leading the technical design of complex AI/ML platforms or datadriven products.

AI Platform & ML Architecture

In-depth knowledge of AI/ML solution design – Large Language Models (LLMs), multi-model orchestration, agent frameworks, and vector databases for embedding storage and semantic search.

Cloud-Native Engineering

Proven cloud-native engineering on AWS and/or Azure using Docker and Kubernetes; familiarity with hybrid-cloud / on-premises integration for sensitive workloads.

MLOps, CI/CD & Observability

Strong MLOps and DevOps practice – CI/CD pipelines for model deployment, and observability with ELK and Grafana (latency, drift, accuracy).

API Management & Secure Gateway Design

Expertise in API gateway and secure LLM-gateway design – centralised key management, request logging, throttling, JWT/OAuth, rate limiting and multi-tenant management.

Enterprise Integration (MCP & Connectors)

Enterprise integration experience using Model Context Protocol (MCP) or similar patterns to fetch enterprise data in a governed way.

Data Governance & Privacy Engineering

Solid data governance and privacy engineering – data quality, dataset versioning, PII handling, anonymisation/tokenisation, and regulatory compliance.

Explainable & Responsible AI

Explainable and responsible AI expertise (SHAP/LIME, fairness and bias evaluation) aligned to ISO 42001 and the NIST AI Risk Management Framework.

Security & Compliance

Strong security and compliance grounding – SOC 2, encryption in transit and at rest, zerodata-retention enforcement, and vendor risk assessment.

Certifications