Data Scientist

• Mar 2022 - Present

Data Scientist

• Jun 2021 - Mar 2022

• Developed a full-stack mobile application that allows the user to scan and annotate a home, and displays an automatically generated 2D floorplan and 3D Virtual Reality (VR) view supported by A-Frame• Built AR sessions in Swift using ARKit and SceneKit; designed and implemented user interfaces with SwiftUI• Implemented API services with Flask that connect the mobile app with Firebase and AWS S3 storage• Designed and built the floorplan algorithm that renders professional-looking floorplans with Shapely in Python• Productionized end-to-end NLP models to analyze NPS (net promoter score) response for better customer satisfaction• Worked closely with established engineers, managed an intern team of three and assisted in project planning

Data Scientist

• Jan 2021 - May 2021

Graduate Student Researcher

• Jan 2021 - Jan 2022

• Evaluated RNNG(rammars) and the GPT2 models’ ability in learning long-term grammatical dependencies • Trained NLP models on gigabyte-size Mandarin datasets in Python and built syntactic test suites • Designed and implemented an Auto-Maze experiment to study human’s processing difficulty on Garden-Path effect as a baseline for language models on Ibex Farm, and generated distractor materials using PyTorch • Evaluated RNNG(rammars) and the GPT2 models’ ability in learning long-term grammatical dependencies • Trained NLP models on gigabyte-size Mandarin datasets in Python and built syntactic test suites • Designed and implemented an Auto-Maze experiment to study human’s processing difficulty on Garden-Path effect as a baseline for language models on Ibex Farm, and generated distractor materials using PyTorch

Data Science Intern

• Jun 2020 - Aug 2020

• Segmented furniture objects from room images using Detectron2, linked them to the most similar-looking products with a KNN model based on the ResNet embedding, color, and shape context, and built an API • Built a hierarchical n-gram template model in Python to generate real estate listing description templates, incorporating Google Maps API for nearby places search and LDA for topic modeling • Led a team of 7 and won the Grand Prize and “Best Tech Demo” Prize at the “REX Intern Hackathon Challenge” with an inventory management project, created a manager UI and an agent UI that can track the inventory and handle restock requests efficiently, saving 70% time than managing the Google Forms

Undergraduate Student Researcher

• Sep 2018 - Aug 2019

• Investigated factors that affect second language learning including word frequency and semantic relationships • Cleaned the dataset with text data wrangling skills including regular expression techniques in R • Built pipelines in Python to do machine translation and calculate translation distances • Conducted several different types of mixed-effects regression analyses in R and used the ggplot2 package to improve data visualization • Investigated factors that affect second language learning including word frequency and semantic relationships • Cleaned the dataset with text data wrangling skills including regular expression techniques in R • Built pipelines in Python to do machine translation and calculate translation distances • Conducted several different types of mixed-effects regression analyses in R and used the ggplot2 package to improve data visualization

Financial Analyst

• Jun 2017 - Aug 2017

• Used R to analyze the advantages and disadvantages of different financial products provided in ICBC; • Collaborated closely with professional financial consultants to identify customer needs and demands • Used R to analyze the advantages and disadvantages of different financial products provided in ICBC; • Collaborated closely with professional financial consultants to identify customer needs and demands