• Constructed time series models and performed explanatory analysis on 4k+ webinar user data to understand webinars’ attendees and its impact on usage frequency and depth• Provided recommendation for webinar based on the analysis results, including new advanced webinar targeting experienced users, more group activities during session, and follow-up email for users who missed the webinar• Created dashboards based to display real-time business metrics related to distinctive areas of interests within customer experience by using Tableau and Preset• Developed explanatory models for product training to help CSMs identify champion and Implementation team drive product adoption• Created and maintained ELTs by writing custom SQL queries, validating data against source applications, and making modifications according to new requirements through DBT

Teach Assistant of BIOST 310: Biostatistics For The Health Sciences, supervised by Professor Allison Meisner- Held discussion including the courser material review and homework Q&A- Held office hour- Provided homework answer keys and graded homework

• Improved the LECIF model by increasing the number of species and tissue types for efficient gene alignment.• Performed a deep learning model on 3 species with multiple tissues to study species conservation at the genomic level.• Built and trained an ensemble of neural networks using 1M+ training functional genomic annotations in Python.• Used ROC curve with 100,000 testing functional genomic annotations, model achieving 85% accuracy.• Performed the new model to study species conservation at the genomic level for pig, mouse, and human species.• Performed a neural network method on 3 species with 4 tissues to study species conservation at the genomic level.

• Built functions to compare the performances of three selected nonparametric methods: MCM, MMCM, and a new model from the lab, on 70,377 T cell data extracted from three locations: normal tissue cells, blood cells and tumor cells for 86 pathways.• Wrote a Python package using API to access genetic code information for each pathway on the KEGG website.• Performed data visualization of the results via ggplot2• Concluded the MMCM model has a better performance in differentiating T cells extracted from different locations.