MBBS and machine learning engineer building end-to-end clinical systems across time-series modelling and healthcare NLP

Currently building out the ML research pipeline at RadNomics, involving large-scale clinical dataset augmentation, unsupervised radiology report generation, proprietary LLM benchmarking, and open-model LoRA fine-tuning

Python · PyTorch · Hugging Face · FastAPI · Docker · Google Cloud Run · GitHub Actions

• Hybrid clinical NLP pipeline generating structured entity outputs from adult ICU progress notes
• Implemented regex-based extraction schemas for recall-focused extraction of 3 clinical entity types
• Fine-tuned and threshold-tuned a BioClinicalBERT classifier on 1000+ manually annotated entities for validation
• Processed 160,000+ ICU notes across 30,000+ stays, extracting 780,000+ structured clinical entities
• Improved validation precision by 45.9% and reduced false positives by 83.3% relative to the rule-only baseline
• Containerised inference service with FastAPI/Docker, deployed on Google Cloud Run with GitHub Actions CI/CD

Live API · Repository · Zenodo DOI

Python · PyTorch · LightGBM · Scikit-learn · SHAP

• Dual-architecture ICU early warning system combining a Temporal CNN (TCN) and LightGBM for NEWS2-derived deterioration prediction across 3 clinical risk dimensions
• Engineered 171 timestamp-level features and 40 aggregated patient-level features across 8 vital parameters, from 70,000+ extracted time-series observations over 140 ICU stays
• TCN improved acute-event AUC by 9% over baseline; LightGBM reduced Brier score by 68% and RMSE by 48% for prolonged-risk prediction
• Implemented clinician-interpretable SHAP and temporal saliency analysis for feature contribution insights

Repository · Zenodo DOI

Applied Machine Learning Engineer @ RadNomics Ltd

• Built radiology data augmentation pipeline from 2.3 million MIMIC-IV reports, producing 15.6 million supervised reconstruction pairs and a 7,000-task evaluation set across seven controlled transformations
• Implemented scalable LLM evaluation framework across 9 proprietary and open models, analysing 63,000 validated generations using textual, semantic, radiology-aware, and operational metrics
• Developed reproducible research within private GCP/GKE infrastructure environment, presenting core research insights to founding team

• MSc, Computer Science with Artificial Intelligence @ City St George’s, University of London
• MBBS, Medicine @ Norwich Medical School, University of East Anglia

• Machine Learning: PyTorch, Scikit-learn, LightGBM, Hugging Face Transformers, PEFT/LoRA, vLLM
• DevOps: Google Cloud Platform (GKE, Cloud Run), Kubernetes, Docker, FastAPI, GitHub Actions (CI/CD)
• Data & Engineering: Python, Pandas, NumPy, SQL (PostgreSQL/MySQL), Git/GitHub