Nottingham, England, United Kingdom
SUMMARY Quantitative Scientist & ML Engineer with a foundation in enterprise software engineering (Siemens) and over a decade of high-impact research at Oxford, KCL, and Manchester. I bridge the gap between rigorous scientific methodology and production-ready AI. With 26 peer-reviewed publications (18 first-author), I specialise in building end-to-end ML pipelines that remain statistically sound in high-stakes environments, from clinical imaging to financial analytics. TECHNICAL SPECIALISMS Production ML: Python (PyTorch, Scikit-learn, XGBoost), FastAPI, Streamlit, and REST APIs. Statistical Rigour: Small-sample inference, cross-validation, and longitudinal 4D data pipelines. Collaborative MLOps: Integrated CI/CD pipelines (Bitbucket, GitHub Actions) and foundational Cloud-native AI (AWS Bedrock, SageMaker). Domain Expertise: Medical Imaging (DICOM/NIfTI), Pharmacokinetics, and Quantitative Finance EDA. TECHNICAL TOOLKIT Languages: Python (Pandas, NumPy, Pytest), R, SQL (MySQL, Oracle), MATLAB, Java (J2EE). MLOps & Tools: MLflow, Optuna, Docker (Foundational), Git, Jira (Agile/Kanban). Scientific Software: Pmod, Amide, ImageJ/Fiji, Hermes, IRTK. WHAT I BRING I am seeking DS/ML/AI roles in the UK (Finance, Healthcare, Life Sciences) where I can apply rigorous data methodology to drive business-critical decisions. I offer the rare combination of a Software Engineer’s foundation and a Scientist’s analytical depth. TECHNICAL PORTFOLIO: https://github.com/Lua-Matlab-Python-R-J2EE
- Engineering Portfolio (MLOps & Full-Stack): Developed end-to-end Python applications to modernise software engineering practices (CI/CD, FastAPI, MySQL). - - Computer Vision Pipeline: Built an automated structural vehicle damage classifier using a custom CNN and pre-trained transfer learning architectures (MobileNetV3/EfficientNet/ResNet) for 2,300 images; deployed via an asynchronous FastAPI backend and a Streamlit dashboard, achieving 75.69% test accuracy. https://car-damage-classifier-v1.streamlit.app/ - - Predictive Systems: Built a healthcare premium prediction engine using a dual-model regression strategy (XGBoost/Scikit-learn) for 50k records; deployed via FastAPI and Streamlit with automated GitHub Actions. https://ml-based-premium-prediction-v1.streamlit.app/ - - Data Engineering: Engineered a full-stack personal finance tracker with MySQL relational database integration and real-time analytics. - - Market Intelligence: Conducted a market viability study for a 50k-record banking dataset, applying A/B Testing and segmentation to define risk-aware product strategies. - Business Analytics: Performed hospitality revenue optimisation for a 150k transaction dataset, identifying key growth drivers and translating complex trends into stakeholder-ready strategies. - Continuing Research: Published 3 peer-reviewed scientific papers in Radiation Oncology (extending Manchester research datasets) while managing a planned career break for primary caregiving. - SQL Systems Reference: Authored a comprehensive technical reference for relational database management, covering complex joins, aggregations, and schema optimisation.
-Predictive Modelling: Engineered and validated predictive models for clinical outcomes (late toxicity) using a multi-centre longitudinal dataset of 1,808 patients. -Geometric Data Engineering: Designed and implemented 3D-to-2D spatial transformation pipelines (cylindrical/spherical mapping) to standardise radiotherapy dose distributions for downstream ML analysis. -Statistical Frameworks: Led three core analytical workstreams applying Propensity Score Matching, Permutation Testing, and Hypothesis Testing to drive clinical decision-making. - - Workstream 1 (Sensitivity Optimisation): Developed a baseline adjustment methodology that significantly increased model sensitivity by optimizing event-count accounting in dose-response curves. - - Workstream 2 (Spatial Feature Engineering): Automated the detection of spatial sub-regions linked to outcome variations, quantifying the impact of hyperparameter choices on feature stability. - - Workstream 3 (Reliability & Scalability): Performed inter-centre reproducibility analysis using harmonised modelling workflows to benchmark independent mapping techniques across multi-site data. - Leadership & Communication: Presented complex 3D/4D visualisations to international stakeholders; supervised junior researchers; and completed Stanford University genomics coursework to integrate bioinformatics into quantitative pipelines.
- Product & Technical Leadership: Directed the full web development lifecycle for an LMS start-up using Agile (Jira/Kanban). Managed a multidisciplinary team of 6 (developers and content specialists). - Infrastructure & Architecture: Oversaw strategic technical decisions including stack selection (Laravel/PHP), cloud infrastructure (AWS/DigitalOcean), and payment gateway integration (Stripe). - Compliance & Data Governance: Spearheaded the company’s GDPR compliance framework, ensuring data privacy and security standards were met across all web-based operations. - Concurrent Quantitative Research: Directed independent research workstreams leading to 8 peer-reviewed publications. Served as an expert peer reviewer for international scientific journals. - Educational Consulting: Designed and delivered advanced university-level modules in Applied Research & Statistics, Calculus, and Stochastic Processes. Provided specialist tutoring in high-level Mathematics. - Scientific Communication: Provided executive editing and review for academic manuscripts, grant proposals, and professional CVs across global research disciplines.
- Cross-Institutional Collaboration: Partnered with leading London institutions (KCL / UCL) to deliver and publish two high-impact scientific papers, demonstrating the ability to manage complex, multi-site research data. - Data Analysis & Research Engineering: Utilised Python and R to build analytical scripts for data processing, statistical modelling, and insight generation to drive peer-reviewed research outputs. - Adaptive Problem Solving: Independently maintained high research productivity and technical upskilling during the COVID-19 pandemic, successfully pivoting to remote collaborative workflows and self-directed project delivery.
- Data Governance & Access: Authored the technical research proposal required to secure access to the National Survey of Health & Development (NSHD), managing the analysis of 1,500 subjects. - Data Wrangling & Processing: Performed extensive data cleaning and preprocessing in R, including missing value imputation and outlier detection, to ensure the integrity of aggregated and subgroup datasets. - Model Validation Framework: Engineered a rigorous statistical benchmarking pipeline to evaluate 10 predictive mathematical models against clinical gold standards. - Advanced Statistical Inference: Applied high-level tests including Shapiro–Wilk (normality), Pearson Correlation, and F-tests to quantify and compare model performance. - Agreement Analysis: Utilised Bland–Altman plots (bias, 95% limits of agreement, and slope analysis) and confidence intervals to identify and validate the most accurate predictive model for clinical application.