Zhi Ye

University of Washington, Information School & Department of Geography

Research Assistant Product Manager at UW Center for an Informed Public

• Led a team of 4 in an Agile environment to develop a Natural Language Processing product that assists researchers in tracking US election mis- and disinformation from social media platforms like Twitter.
• Utilized Python to create 7 features based on the voice of the customer (VOC). These features include sentiment analysis, search, standardizing timezones, popular tweets, and filtering (date, time, retweets).
• Drove the implementation of product features by identifying the problem, defining requirements, performing market and user research, and working with design and engineering to deliver upon those requirements.
• Conducted 60-minute Zoom interviews with 6 CIP qualitative researchers to identify user’s unmet needs and collect information about their research regarding current methods used, time spent on tasks, and pain points.
• Engaged with key stakeholders to gather input and garner support for product vision, strategy, and roadmap.
• Created and managed backlog of work through the product’s life cycle and a comprehensive 14-page handoff documentation detailing project overview, research, design direction, product features, testing, metrics, etc.
• Managed communication across 8 CIP clients and 2 UW mentors to schedule and execute weekly meetings to discuss project deliverables (concept-MVP-final product), features, enhancements, roadblocks, and deadlines.
• Spearheaded product presentation and demo at CIP’s research symposium to an audience of 50+ including CIP clients, CIP staff, UW faculty, undergraduate students, graduate students, PhDs, and researchers.

Waiter at Kim Bowl Restaurant & Lounge

• Managed the entire guest experience as the primary waiter of the restaurant. Duties included providing exceptional customer service, offering tailored menu suggestions, and assisting with customer requests, questions, and concerns.
• Spearheaded a seamless end-to-end guest experience by consistently being aware of customer needs and wants—inclusive of timely check-in, executing additional requests, and bill transactions.
• Worked closely with kitchen staff and head chef to make menu suggestions or adjustments based on customer feedback.
• Performed tasks in accordance with training, safe food handling procedures, and customer dietary restrictions.
• Maintained cleanliness of all dining spaces, checkout counter, restaurant entrance to upkeep a professional appearance.
• Contributed to a supportive and inclusive team environment by engaging in dialogue on work process improvement, improving guest experience, implementing customer feedback, and upsell opportunities.

Cheffy, a Social Cooking App [Miro] [Figma]

• Conducted extensive market and user research by analyzing 11 cooking/social media competitors, interviewed 12 individuals, and surveyed 20+ people to inform target demographics, user needs, design decisions and features.
• Performed usability testing with 7 individuals and 12 participants from usertesting.com to evaluate the functionality of the product and discover new opportunities for improvements based on user behavior, needs, and expectations.
• Designed high-fidelity prototypes in Figma with 30 screens filled with interactions and animations for mobile devices to gain detailed feedback from usability testing by effectively implementing flows, visuals, navigation, and engagement.

Predicting Adult Income That Will Exceed $50,000 per Year [Github]

• Employed Plotly in Python to create 3 interactive data visualizations featuring stacked and grouped bar graphs to showcase and examine income inequality across demographic factors such as gender, race, occupation, and education.
• Analyzed, cleaned, and performed supervised machine learning by comparing models such as decision tree and k-nearest neighbors on the Adult Income Dataset using Python. Dataset was split into 70% training and 30% testing.
• Computed testing accuracy of 0.785 with DecisionTreeClassifier and 0.33 with KNeighborsClassifier for adults with income greater than $50,000 to determine best performing model with given dataset.
• Created 9 unit tests to ensure the quality of my code, assist with debugging, and provide thorough documentation.
• Consolidated research, methodologies, visualizations, and results to create a 10-page detailed report of my findings.

Healthcare Database [Github]

• Developed a secure, scalable, and reliable healthcare database with 29 tables using data definition and data manipulation language in SQL Server that allows medical professionals to store and query patient information.
• Utilized stored procedures and wrappers to create 800,000 synthetic transactions to test database performance.
• Produced 10 business rules along with 10 check constraints to implement business policies and preserve data integrity.
• Designed an entity relationship diagram (ERD) in Lucidchart that includes 29 entities with various attributes, primary/foreign keys, and relationships (e.g., one-to-many) to model the data that will be stored within the database.

Spatial & Temporal Analysis of U.S. Life Expectancy [Tableau Public]

• Utilized Tableau Public to create a story that communicates male and female life expectancy from 1985-2010 at 3 spatial scales: national, state, and county.
• Created 8 story points: 6 interactive choropleth maps and 2 bar graphs with detailed insights to help drive the primary narrative of my analysis.

Wisconsin Breast Cancer Classification [Github]

• Analyzed and cleaned the Wisconsin Breast Cancer Dataset.
• Extracted specific variables such as diagnosis, concpoints.mean, fracdim.mean, etc. to analyze numerous logistic regression models.
• Computed accuracy and F-score on models to predict whether the diagnosis is malignant or benign.