Architected internal MLOps platform supporting 2,000+ model training experiments with hyperparameter tracking
Designed Agentic LLM system with tool use to answer marketing data analysis questions and extract KPIs
Deployed 50+ models for prescriptions forecasting, owning ETL optimization, drift monitoring and retraining
Peptone
Machine Learning Research Engineer
Fine-tuned protein language models to run semantic search on antibody libraries for antigen targets
Screened drug hits across 20M ligands with pLMs and GNNs distributed over 40 GPUs on Docker/Kubernetes
Improved a biological LLM-chatbot by 10% with Retrieval-Augmented Generation
Perspectum
Senior Data Scientist
Scaled MLOps capabilities to over 100 models using AWS Batch/ECS and MLflow
Trained large 3D vision models on 1 GPU via DeepSpeed with acceleration and gradient checkpointing
Fine-tuned diffusion models using LoRA on synthetic MRI images of rare anatomy for data augmentation
ONI (Oxford Nanoimaging)
Deep Learning Engineer
Used self-supervised image-denoising PyTorch models to decrease laser power for microscopy by 20%
Applied transfer learning to reduce retraining data needs by 80% across new imaging tasks
MasonBreese
Consultant Data Analyst
Built ETL pipelines in SQL for 60,000 customers and over £1bn in retail bank assets
Identified churn-risk customers with tree models using Scikit-Learn
Education
University of Warwick
MMath, Master of Mathematics (1st)
Google Cloud Certified
Professional Data Engineer
AWS
Certified Solutions Architect Associate
Publications
Improving Inverse Folding models at Protein Stability Prediction without additional Training or DataNeurIPS 2024 Machine Learning Structural Biology WorkshopOptimizing protein language models with Sentence TransformersNeurIPS 2023 Machine Learning Structural Biology Workshop