Yuxing Fang Email & Phone Number

London, England, United Kingdom

• Successfully mapped millions of clinical data points to approximately 400 concepts, leveraging word embedding methods for robust semantic standardization in NLP.
• Crafted a hybrid LLAMA2 rule-based model that streamlined knowledge extraction, boosting processing speed by 1000x.
• Launched Regex-Builder, an integrated development suite tailored for researchers without AI expertise, featuring an intuitive interface with templates, validation, and a real-time sandbox tailored for patient data.

Oxford, England, United Kingdom

• Engineered an NLP training framework for patient EHRs, incorporating NER and classification models, and coding for deployment, processed over 4 million patient records.
• Refined BERT and RoBERTa models, halving execution time, cutting operational expenses, and boosting accuracy via hyperparameter optimization.

• Developed DNN-based models for real-time decoding of semantic concepts from direct brain recordings, enhancing model interpretability and computation transparency.
• Implemented Representational Similarity Analysis to align BERT's sentence interpretation with the representational structure of brain activities.
• Employed DNN-HMM hybrid models with topic modeling to map millisecond-level semantic and phonological representations in the cortex, presenting findings in a detailed poster.

Beijing City, China

• Developed an SVM-based method for testing semantic representations in white matter connectivity. Findings demonstrated that white matter connections and grey matter regions encode different types of semantic.
• Full stack developer (architecture design, hardware installation, software and network configuration, HCP software development, and daily management and maintenance)
• Maintained two systems for clinical imaging processing (512 CPU cores / and ~1000 CPU cores systems).

Cambridge, USA

Doctor Of Philosophy - Phd, Psychology

Bachelor'S Degree, Horticultural Science