• Fine-tuned LLMs for domain-specific use cases using HuggingFace datasets with 10000+ data points.• Deployed optimized models with AWS SageMaker and S3 for improved model accessibility.

• Achieved 76.3% validation accuracy with an MLP model engineered using PyTorch and scikit-learn.• Automated data collection with Python scripts executed in a Linux-based physics simulation environment.• Experimented with dimensionality reduction techniques such as CNNs and PCA to optimise performance.

• Cleaned and consolidated UK government school data from spreadsheets and inspection reports.• Improved employee working efficiency through developing Python data processing tools to streamline Excel spreadsheet workflows.

• Refactored and enhanced a web application to host analysis tools, enabling NHS practitioners to efficiently address patient concerns from phone conversation recordings.• Deployed the webapp onto an NHS Nginx server. • Collaborated with stakeholders to formulate requirements and user stories using Trello dashboards.