🎁 Before you apply, rehearse this interview. Create your free WorkMundi account and get an Interview Training on HelpsYouSpeak — no cost, no card. I want my training →
Responsibilities • Independently assess AI/ML/data science model purpose, assumptions, features, data inputs, and logical soundness. • Evaluate feature engineering, data quality, and detect issues such as leakage or mis-specified inputs. • Evaluate model performance using suitable metrics, diagnostic tests, and validation methodologies. • Assess stability, robustness, sensitivity analysis, susceptibility to adversarial attacks and model or concept drift. • Apply model explainability methods such as SHAP, LIME and other interpretability techniques. • Produce comprehensive, well-reasoned Model Validation Reports. • Evaluate AI/ML models, LLMs, retrieval-augmented systems, agentic workflows, and prompt-engineering methods. • Ensure validation standards align with Responsible AI principles including fairness, transparency, and robustness. • Collaborate with data scientists and model developers across business and functional teams to understand modelling intent, design rationale, and underlying assumptions. • Contribute to exploratory AI/ML proof ‑ of ‑ concept (POC) initiatives to deepen technical understanding, enhance validation methods, and support innovation within DMO. Requirements • Preferably a postgraduate degree in Data Science, Statistics, Mathematics, Analytics, Computer Science, or quantitative discipline. • At least 4 years of hands ‑ on experience in model development, model validation, quantitative analytics, or AI/ML evaluation within financial institutions or similarly regulated environments. • Strong theoretical and practical knowledge of machine learning, AI, statistical models, and model validation techniques. • Strong understanding of feature engineering, feature selection, and data quality checks. • Proficiency in evaluating model performance and diagnostics across statistical, ML, and AI models. • Understanding of explainability techniques, including SHAP, LIME, and other model interpretation methods. • Analytical skills to identify modelling weaknesses, design flaws, and performance gaps. • Strong reporting skills to produce high-quality validation deliverables. • Familiarity with Responsible AI concepts such as fairness, transparency, and robustness.