Bing, Whole Page, WebXT

Working on the base layer data for answer and knowledge card in Bing.

• Increased data coverage and quality by onboarding data, crowdsourcing curation, and rebuilding the knowledge graph. Designed and implemented scalable data pipelines to replace obsolete or absent data with new resources. Optimized and generalized prompts with Large language models(LLMs) for data parsing, sourcing and validation. Led projects by coordinating multiple teams' requests.

• Developed a data coverage and quality measurement system for leadership, PMs, and data owners to facilitate data-driven business decisions. Dashboards effectively visualize large amounts of data with tables and graphs. Led whole project across multiple stakeholders, including data owners, PMs, leadership, crowdsourcing teams, and development vendors.

• Leveraged LLMs, crowd sourcing, A/B testing, etc.

Worked on AWS services -

• Utilized cloud services to design and implement scalable system – data pipelines, software UI changes, environment set up using Active Directory authentication across OS, etc.

• Mentored a 2021 summer intern, finalizing and deploying their work to production.

One Lambda, Transplant Diagnostics

Worked on overall data system for Transplant Diagnostics Kits.

• Designed and built NGS(Next Generation Sequencing) data process integration system (Pipeline) with Python and C#

• Developed classification model to predict primer set amplifying status using Python pandas, numpy and scikit-learn

• Designed and developed internal HLA sequence library software using Python and C++

• Updated internal database regularly by collecting, cleaning and validating data from various sources

• Developed multiple Python programs to improve data processing (Workflow automation)

• Maintained multiple software written by R, Python

• Implemented data conversion software to support laboratory scientists with C#

• Generated and updated technical specific documents for customer-facing products

Worked for Bioinformatics research group to help professors from Biology domain.

• Automated multi-step analysis pipelines (Gene sequencing tools) on Linux environments with Python

• Generated heatmap, scatter, cluster plots, etc. to visualize data with R

• Revised scripts from existing R packages (Cell Ranger, Monocle, Seurat, etc.) to handle varying biologist requests

• Analyzed NGS data and single cell RNA-seq data from wet laboratories with R and Python.