Chayenne Mosk

Led development of AI-powered knowledge discovery solutions, including RAG, semantic search, and LLM-based information retrieval tools to accelerate research and decision-making Evaluated generative AI approaches for research and knowledge-management use cases, benchmarking retrieval quality, groundedness, and operational suitability to identify effective business applications Investigated agentic AI workflows, prompt-engineering strategies, and orchestration approaches to identify practical, evidence-based applications within a regulated environment

Data Scientist | Co-lead Statistics Network - UK National Audit Office (NAO) - London, UK

(2024-05)

• Led development of AI-powered knowledge discovery solutions, including RAG, semantic search, and LLM-based information retrieval tools to accelerate research and decision-making
• Evaluated generative AI approaches for research and knowledge-management use cases, benchmarking retrieval quality, groundedness, and operational suitability to identify effective business applications
• Investigated agentic AI workflows, prompt-engineering strategies, and orchestration approaches to identify practical, evidence-based applications within a regulated environment
• Designed and implemented embedding-based search pipelines using Azure OpenAI and Microsoft Foundry, supporting migration to modern enterprise AI infrastructure
• Built and validated credit-risk and forecasting models across £7bn+ government-backed portfolios using statistical modelling, machine learning, and multi-scenario macroeconomic forecasting
• Communicated complex analytical findings, model limitations, and AI performance metrics to technical and non-technical stakeholders, supporting operational and strategic decision-making
• Co-lead the Statistics Network, driving adoption of emerging statistical, machine learning, and AI methods through knowledge-sharing and training across the organisation

Data Scientist - Belmana | Data Analytics Consultancy - London, UK

(2023-10 - 2024-05)

• Developed Python- and SQL-based data processing workflows for multi-million-record datasets, reducing manual effort and improving analytical efficiency
• Conducted causal impact and opportunity-sizing analyses using PSM and difference-in-differences, directly informing funding allocation decisions and improving SME acquisition efficiency by 15%

Quantitative Research Consultant - Deutsche Bank - London, UK

(2022-09 - 2023-09)

• Built and optimised ML models (LSTM, XGBoost, RF, SVM) on 15 years of FX data across 17 global markets, improving next-day forecasting accuracy by 5%
• Developed model interpretability frameworks (SHAP, LIME, PFI) to explain complex ML predictions, increasing stakeholder trust and adoption of model-driven insights
• Designed and backtested model-driven trading strategies using forecast signals and risk metrics (PnL, Sharpe ratio), achieving a 7% increase in risk-adjusted returns

Data Science Intern - ING Bank - Amsterdam, NL

(2022-01 - 2022-03)

• Built ARIMA forecasting models with macroeconomic drivers to project ING's balance sheet to 2050

MSc - Data Science - London School of Economics and Political Science (LSE) (2022-09 - 2023-08)

BSc (Hons) - Econometrics & Data Science - Vrije Universiteit (VU) (2019-09 - 2022-08)