AI Data Scientist

About the Role:
The AI Data Scientist is responsible for translating business and solution requirements into clear data, feature engineering, and modeling approaches for selected AI use cases. The role develops, evaluates, and improves machine learning and statistical models, ensuring they are accurate, robust, explainable, and aligned with measurable business outcomes.

Working under the direction of the AI Project Portfolio Lead, the role collaborates with Applied AI Engineers, AI Champions, business teams, Data, Technology, Architecture, Risk, and Information Security to prepare data, benchmark model options, and support models through testing, deployment, and ongoing monitoring. The role also helps establish reusable data science practices, evaluation methods, and model standards across the AI transformation program.

Key Accountabilities

• Translate business requirements and approved solution designs into clear data, feature, analytical, and modeling requirements.
• Work with business stakeholders, AI Champions, and solution teams to define target variables, outcomes, evaluation criteria, and model success measures.
• Explore, analyze, clean, and curate structured and unstructured data required for selected AI use cases.
• Identify data quality, completeness, lineage, representativeness, leakage, and bias issues, and coordinate remediation with relevant data owners.
• Design and engineer relevant features, variables, and datasets for model development, testing, and validation.
• Build, train, tune, and evaluate machine learning, statistical, and predictive models using appropriate techniques and technologies.
• Benchmark internally developed models against baseline approaches, third-party models, vendor solutions, and alternative analytical methods.
• Assess model accuracy, precision, recall, robustness, stability, explainability, fairness, latency, and operational suitability.
• Define and maintain appropriate training, validation, and testing methodologies, including representative datasets and performance thresholds.
• Support the evaluation and optimization of generative AI and large language model solutions where model selection, retrieval quality, classification, or output evaluation is required.
• Prepare clear model documentation covering assumptions, methodology, data sources, features, limitations, testing results, and intended use.
• Work with Applied AI Engineers and Technology teams to integrate approved models into applications, workflows, APIs, and production environments.
• Support model deployment, versioning, performance monitoring, drift detection, retraining, and continuous improvement.
• Define monitoring metrics and thresholds to identify deterioration in accuracy, data quality, bias, stability, or business performance.
• Apply approved data governance, privacy, security, model risk, and responsible AI controls throughout the model lifecycle.
• Support model validation, governance, Information Security, Risk, and audit reviews by providing clear evidence, documentation, and technical explanations.
• Develop reusable data preparation pipelines, evaluation frameworks, feature engineering components, and model development standards.
• Provide clear updates, insights, and recommendations to the AI Solution Lead and relevant stakeholders on model performance, data readiness, risks, and improvement opportunities.
• Contribute to a collaborative, accountable, and performance-driven culture focused on measurable outcomes and responsible use of data and AI.

Experience, Skills & Success Profile

Experience & Qualifications

• Relevant hands-on experience in data science, machine learning, advanced analytics, predictive modeling, statistical analysis, or a related role.
• Bachelor’s degree in data science, computer science, statistics, mathematics, engineering, econometrics, or a related quantitative discipline, or equivalent professional experience.
• A master’s degree or advanced qualification in a relevant quantitative or technical discipline is advantageous.
• Demonstrated experience developing, evaluating, and deploying machine learning or statistical models for practical business use cases.
• Strong programming skills in Python, R, SQL, or other relevant data science and analytics technologies.
• Experience using machine learning frameworks and libraries such as scikit-learn, PyTorch, TensorFlow, XGBoost, or comparable technologies.
• Strong experience in data exploration, cleaning, feature engineering, model selection, tuning, validation, and performance evaluation.
• Understanding of supervised and unsupervised learning, classification, regression, forecasting, clustering, anomaly detection, and optimization techniques.
• Familiarity with generative AI, large language models, embeddings, retrieval-augmented generation, and model evaluation is highly desirable.
• Experience with cloud data and AI platforms, model development environments, notebooks, version control, and data pipelines is preferred.
• Understanding of model monitoring, drift detection, retraining, model lifecycle management, and MLOps practices.
• Experience evaluating third-party models, vendor solutions, or externally sourced analytical capabilities is advantageous.
• Familiarity with data governance, model risk management, responsible AI, explainability, privacy, and bias assessment.
• Experience within financial services, fintech, payments, fraud, risk, customer analytics, or another regulated and data-intensive environment is preferred.
• Ability to perform with pace, accuracy, and sound judgment in a fast-moving, commercially focused environment.

Skills & Competencies

• Demonstrates strong analytical, statistical, and quantitative problem-solving capability.
• Translates business questions into practical modeling approaches, measurable outcomes, and clear evaluation methods.
• Selects appropriate algorithms and techniques based on the business problem, data characteristics, and operational requirements.
• Understands the strengths, limitations, and risks of machine learning, statistical, and generative AI models.
• Applies rigorous experimentation, validation, and benchmarking practices rather than relying on model performance in isolation.
• Identifies data quality, leakage, bias, representativeness, and reliability issues early in the development process.
• Communicates model results, assumptions, limitations, and business implications clearly to technical and nontechnical audiences.
• Uses data, evidence, and structured analysis to challenge assumptions and support better decisions.
• Maintains strong attention to detail across datasets, code, features, experiments, documentation, and model outputs.
• Writes clear, maintainable, and reproducible code and applies appropriate version control and development standards.
• Balances technical sophistication with interpretability, scalability, operational practicality, and business value.
• Collaborates effectively with business stakeholders, solution leads, engineers, data teams, architects, and control functions.
• Plans and prioritizes effectively across multiple models, experiments, datasets, and delivery dependencies.
• Solves problems proactively with a practical, structured, and solution-oriented mindset.
• Takes clear ownership of model quality, data issues, testing results, and technical risks.
• Learns new techniques, platforms, and modeling approaches continuously in a rapidly evolving AI environment.

What Success Looks Like in This Role

• Business and solution requirements are translated into clear, relevant, and technically sound data and modeling approaches.
• Data quality, completeness, bias, and readiness issues are identified early enough to avoid preventable delivery delays or unreliable outcomes.
• Models are developed using appropriate methodologies and meet agreed accuracy, robustness, explainability, and performance expectations.
• Model choices are supported by clear benchmarking against baselines, alternative approaches, and relevant third-party solutions.
• Evaluation methods, testing datasets, and performance thresholds are well defined, repeatable, and aligned with business outcomes.
• Model assumptions, limitations, risks, and intended use are documented clearly and understood by relevant stakeholders.
• Approved models integrate effectively into enterprise applications, workflows, and production environments.
• Model performance, data quality, and drift are monitored consistently after deployment, with clear triggers for intervention or retraining.
• Responsible AI, privacy, security, fairness, and model governance requirements are embedded throughout the model lifecycle.
• Reusable data pipelines, evaluation frameworks, and modeling components improve the speed and quality of future delivery.
• Stakeholders view the role as analytically rigorous, commercially aware, collaborative, and focused on solving meaningful business problems.
• Models contribute measurable improvements in revenue, efficiency, forecasting, risk management, customer experience, or operational decision-making.
• The role contributes positively to technical quality, responsible innovation, collaboration, and continuous improvement across the AI transformation program.