Focused on advancing bioinformatics through machine learning and deep learning applications. Developed BreathXplorer, a Python package designed to recognize metabolic features from breath data using unsupervised machine learning techniques. Additionally, created a deep learning generative model utilizing Transformer architecture to predict metabolite structures, offering promising applications in drug discovery.On the same time, I also worked as the system administrator in the lab. I set up on-premise HPC clusters in the laboratory and utilized Docker for environment virtualization for web deployment, big data pipelines, and deep learning.

Developed a health consultation system utilizing Retrieval Augmented Generation (RAG), integrating Large Language Models (LLMs) with a vector database.Engineered the system to provide accurate responses to medical queries, ensuring answers are backed by verified and traceable scientific research for user reliability.

Designed and implemented a large-scale ETL pipeline to curate and process over 500GB of genomics data using Python, Bash, SQL, and HPC environments.Developed a state-of-the-art GBDT model with scikit-learn to accurately predict pathogenic gene mutations.