Job Description

AI-First Solutioning & Architecture (Core)

- Lead end-to-end architecture for AI/ML systems: data features training evaluation deployment monitoring retraining.

- Create and maintain reference architectures for:

- Batch scoring (e.g., nightly risk insights),

- Near-real-time decisioning (fraud, eligibility checks),

- Personalisation (ranking/next-best-action),

- Customer servicing intelligence (triage, routing, summarisation, agent assist).

- Drive build-vs-buy decisions for platforms and tools (ML platform, vector DB, model registry, feature store, orchestration, monitoring).

- Define non-functional requirements: latency, scale, resilience, privacy, explainability, auditability, cost, and operational ownership.

Applied ML Engineering (Hands-on)

- Build and productionise models across:

- Fraud/anomaly detection, entity resolution, risk scoring,

- Credit insights and predictive risk indicators,

- Propensity/CLV/churn models, uplift modelling, recommender/ranking systems,

- Customer servicing classification, intent detection, routing optimisation,

- (Where relevant) LLM-enabled workflows with evaluation + guardrails.

- Design rigorous evaluation strategies:

- Offline metrics (AUC/PR, calibration, cost-based thresholds),

- Backtesting and stability checks,

- Fairness/bias and subgroup performance,

- Online experimentation (A/B, multi-armed bandits where appropriate).

MLOps, Reliability & Observability (Staff expectation)

- Establish ML engineering standards: CI/CD for ML, reproducibility, model registry, lineage, approvals, and release strategies (canary/blue-green).

- Build monitoring for:

- Data quality + drift,

- Model performance decay and calibration drift,

- Pipeline health and SLA/SLO compliance,

- Business KPI impact and guardrail metrics.

- Define retraining triggers and human-in-the-loop workflows where required.

- Ensure secure-by-design practices: secrets management, encryption, IAM, audit logs, and incident runbooks for AI services.

Stakeholder Management & Business Partnership

- Partner with Product, Marketing, Risk, Compliance, Legal, Security, and Operations to:

- Frame problems, define success metrics, and quantify ROI,

- Navigate tradeoffs (accuracy vs latency vs cost vs explainability),

- Align on governance and operational ownership.

- Lead architecture reviews and communicate effectively to both technical and executive audiences.

- Translate consumer outcomes and business needs into measurable AI roadmaps.

Responsible AI, Governance & Privacy (Regulated environment)

- Build Responsible AI into the lifecycle: transparency, fairness, privacy, security, accountability.

- Produce governance artifacts: model cards, data documentation, validation reports, explainability evidence, and audit-ready change logs.

- Ensure UK/EU privacy-aware design practices: data minimisation, retention, lawful basis, access control, and privacy-by-design patterns.

Technical Leadership & Mentorship

- Mentor engineers/data scientists in production ML, architectural thinking, and engineering excellence.

- Establish patterns that scale across teams (templates, libraries, golden paths, reusable services).

- Raise quality across the org through design reviews, standards, and coaching.

What Success Looks Like

- AI initiatives deliver measurable outcomes in ECS UK domains:

- Fraud loss reduction, improved fraud detection recall/precision with low false positives,

- Better consumer credit insights and decision quality with calibrated outcomes,

- Increased marketing conversion/retention via personalisation and uplift,

- Reduced servicing handle time and improved first-contact resolution.

- Teams ship models faster and safer due to platform leverage, strong MLOps, and governance.

- Stakeholders trust and adopt your solutions due to clarity, auditability, and measurable impact.

Required Qualifications

- 12+ years in software/data/ML engineering with 3+ years leading architecture for production ML systems.

- Proven track record shipping multiple ML systems to production with measurable business impact.

- Strong ML foundations: leakage avoidance, calibration, imbalanced learning, evaluation rigor, drift management.

- Deep data expertise: SQL, data modeling, batch + streaming concepts, data quality and lineage.

- Cloud architecture experience (AWS/Azure/GCP) including security, networking, cost, reliability and observability.

- Strong stakeholder leadership: influence without authority, crisp communication, negotiation, and pragmatic delivery.

Preferred / Nice-to-have (Highly valued for ECS UK)

- Fintech experience: fraud, credit risk, identity, payments, AML/KYC, underwriting, collections.

- Real-time decisioning systems and feature stores (low latency constraints).

- Experience with marketing science: uplift, MMM/attribution concepts, propensity models, next-best-action/ranking.

- LLM application experience in regulated settings (RAG, evaluation harness, guardrails, red-teaming, prompt security).

- Familiarity with privacy, compliance, and audit expectations in UK/EU contexts

Qualifications

Qualifications

Typical Tech Stack (Representative not mandatory)

- Languages: Python, SQL (strong) optional: Java/Scala

- ML: scikit-learn, XGBoost/LightGBM, PyTorch/TensorFlow

- Data: Spark/Databricks, Airflow/Dagster/Prefect, dbt (optional)

- Serving: FastAPI, Docker, Kubernetes, model serving frameworks (Seldon/KServe/Triton)

- Streaming: Kafka/Event Hubs/Kinesis

- MLOps: MLflow/W&B, feature store (Feast/Tecton), model monitoring tools

- Observability: OpenTelemetry, Prometheus/Grafana, ELK/Datadog

- Governance: access control, encryption, audit logging, documentation workflows

Example Initiatives You Might Lead

- Fraud detection + entity resolution platform with streaming features and low-latency inference

- Consumer credit insight models with explainability and audit-ready governance

- Personalisation engine (ranking/next-best-action) with experimentation + uplift measurement

- Customer servicing intelligence: triage + routing + agent assist with robust evaluation and guardrails

- Standardised ML deployment golden path improving time-to-production and reliability across ECS UK

Interview Loop (Suggested)

1) ML System Design (end-to-end + tradeoffs)

2) Applied ML Deep Dive (evaluation, drift, calibration, bias/fairness)

3) Coding/Data Exercise (production-quality Python/SQL)

4) Stakeholder/Business Case (ROI, prioritisation, governance)

5) Leadership & Mentoring (influence, standards, decision-making)

Additional Information

Our uniqueness is that we celebrate yours. Experian's culture and people are important differentiators. We take our people agenda very seriously and focus on what matters DEI, work/life balance, development, authenticity, collaboration, wellness, reward & recognition, volunteering... the list goes on. Experian's people first approach is award-winning World's Best Workplaces™ 2024 (Fortune Top 25), Great Place To Work™ in 24 countries, and Glassdoor Best Places to Work 2024 to name a few. Check out Experian Life on social or our Careers Site to understand why.

Experian is proud to be an Equal Opportunity and Affirmative Action employer. Innovation is an important part of Experian's DNA and practices, and our diverse workforce drives our success. Everyone can succeed at Experian and bring their whole self to work, irrespective of their gender, ethnicity, religion, colour, sexuality, physical ability or age. If you have a disability or special need that requires accommodation, please let us know at the earliest opportunity.

Experian Careers - Creating a better tomorrow together

Senior Staff Software Engineer (AI/ML)