Ziwei Gu
Ph.D. Candidate at Harvard University- Claim this Profile
Click to upgrade to our gold package
for the full feature experience.
-
English Native or bilingual proficiency
-
Chinese Native or bilingual proficiency
-
Spanish Limited working proficiency
Topline Score
Bio
Experience
-
Harvard University
-
United States
-
Higher Education
-
700 & Above Employee
-
Ph.D. Candidate
-
Aug 2022 - Present
Research interests: human-computer interaction, machine learning, natural language processing Research interests: human-computer interaction, machine learning, natural language processing
-
-
-
Lyft
-
United States
-
Ground Passenger Transportation
-
700 & Above Employee
-
Data Scientist Intern
-
Jun 2020 - Jul 2020
- Estimated the opportunity and risk of Lyft Family (a new feature of Lyft) by exploring existing rides and payment data - Identified all household users of Lyft and analyzed their unique characteristics in terms of engagement, experience, etc. - Clustered family riders and rides and recommended incentive products targeting each segment of users and use cases - Estimated the opportunity and risk of Lyft Family (a new feature of Lyft) by exploring existing rides and payment data - Identified all household users of Lyft and analyzed their unique characteristics in terms of engagement, experience, etc. - Clustered family riders and rides and recommended incentive products targeting each segment of users and use cases
-
-
-
Cornell Engineering
-
United States
-
Education Administration Programs
-
100 - 200 Employee
-
Teaching Assistant
-
Jan 2020 - May 2020
CS 3410: Computer System Organization and Programming
-
-
Teaching Assistant
-
Aug 2019 - Dec 2019
CS 4780: Machine Learning for Intelligent Systems
-
-
Teaching Assistant
-
Aug 2018 - Aug 2019
CS 2110: Object-Oriented Programming and Data Structures (with Java)
-
-
-
Cornell CIS (Computing and Information Science)
-
Ithaca, New York Area
-
Research Assistant
-
Jan 2019 - May 2020
- Co-author: JN Yan, Z Gu, H Lin, J Rzeszotarski. “Enhancing Sensemaking in Machine Learning Fairness Assessments”. CHI 2020. - Developed an interactive bias exploration dashboard with Flask and Bokeh, helping data analysts reason about the source of bias in their classifiers through feature engineering and shortening their average decision-making time by 22.5 s. - Re-formulated the open information extraction problem as a sequence-to-sequence transduction task and trained a Transformer… Show more - Co-author: JN Yan, Z Gu, H Lin, J Rzeszotarski. “Enhancing Sensemaking in Machine Learning Fairness Assessments”. CHI 2020. - Developed an interactive bias exploration dashboard with Flask and Bokeh, helping data analysts reason about the source of bias in their classifiers through feature engineering and shortening their average decision-making time by 22.5 s. - Re-formulated the open information extraction problem as a sequence-to-sequence transduction task and trained a Transformer encoder-decoder model with PyTorch to extract relational triples from sentences. - Evaluated the model and showed it was on par with the state-of-the-art, but without dependencies on other NLP tools. Show less - Co-author: JN Yan, Z Gu, H Lin, J Rzeszotarski. “Enhancing Sensemaking in Machine Learning Fairness Assessments”. CHI 2020. - Developed an interactive bias exploration dashboard with Flask and Bokeh, helping data analysts reason about the source of bias in their classifiers through feature engineering and shortening their average decision-making time by 22.5 s. - Re-formulated the open information extraction problem as a sequence-to-sequence transduction task and trained a Transformer… Show more - Co-author: JN Yan, Z Gu, H Lin, J Rzeszotarski. “Enhancing Sensemaking in Machine Learning Fairness Assessments”. CHI 2020. - Developed an interactive bias exploration dashboard with Flask and Bokeh, helping data analysts reason about the source of bias in their classifiers through feature engineering and shortening their average decision-making time by 22.5 s. - Re-formulated the open information extraction problem as a sequence-to-sequence transduction task and trained a Transformer encoder-decoder model with PyTorch to extract relational triples from sentences. - Evaluated the model and showed it was on par with the state-of-the-art, but without dependencies on other NLP tools. Show less
-
-
-
Cornell Data Science
-
United States
-
Information Services
-
1 - 100 Employee
-
Project Lead
-
Feb 2018 - Dec 2019
- Led sub-team of 4 in knowledge graph construction from raw text, using a pipelined process of coreference resolution and relation tuple integration - Modeled a virtual traffic network using directed graphs and collected data by running simulations - Improved the network by optimizing traffic objectives with respect to network parameters under various constraints - Led sub-team of 4 in knowledge graph construction from raw text, using a pipelined process of coreference resolution and relation tuple integration - Modeled a virtual traffic network using directed graphs and collected data by running simulations - Improved the network by optimizing traffic objectives with respect to network parameters under various constraints
-
-
Education
-
Cornell University
Bachelor of Arts - BA, Mathematics and Computer Science