Phu Dang

Technical and Leadership Experience

“Links to disclosable code, presentations, and deliverables are available on my LinkedIn profile”

Organization website: Qualcomm Institute

Contributed to developing digital twin platforms for UC San Diego, Chula Vista Bayfront, and Hollywood’s Vinyl District.

• Implemented scalable data integration workflows using blueprints and structs in Unreal Engine 5
• Developed data visualization applications using UE5 objects/instanced static meshes with dynamic spawning, scaling, georeferencing
• Implemented Cesium cartographic polygons, raster overlays, and color blending across material layers for tile coloring
• Built interactive, 3D widgets to display urban data, emphasizing user experience and future VR integration
• Integrated Google Maps API and SketchUp Pro Datasmith Exporter plugin for property and site plan modeling
• Tested frameworks to build twins of real buildings for future placement in the city twins based on photogrammetry scans
• Aggregated a large database of the San Diego-Carlsbad region planning & environmental data for future visualization

Impact report, website

• Manage organizational leadership, budget, marketing, outreach, public relations, and recruitment activities
• Worked with local healthcare non-profits in San Diego on potential avenues for data science to improve operations. Targeting ways to help clients improve efficiency, impact, outreach, client satisfaction, ROI, data systems and use case designs
• Oversee and participate in student data science projects to ensure quality, timely delivery, and meeting clients’ expectations while upkeeping self; facilitating opportunities for students to apply academic training in impactful, real-world scenarios
• Organize correspondence between members, university faculty, department advisors, and non-profit representatives

Lead project teams in developing data-driven solutions to address clients’ operational challenges and objectives. Leverage data workflows ranging from simple to sophisticated, including retrospective, exploratory, and predictive analyses, large language models, and parallel computing with large datasets. Enhanced clients’ understanding of patient programs, donor initiatives, grants performance, business strategies, and investment decisions across over $140 million in combined financial analyses performed.

Advise and collaborate with team members to ensure correct execution, appropriate analysis techniques, and practical, actionable results. Present findings through dashboards, web apps, and slide presentations, emphasizing intuitiveness, practicality, and actionable values that drive informed decisions.

Driven by a commitment to exceptional service quality, providing insights that reflect clients’ past operations while informing recent and future decisions. Achieving a perfect track record of positive client testimonials and satisfaction.

Organization website: WDC

• Advised the ADUGenius project on ADUs as an innovative housing solution, covering financing structures, tax benefits, saving strategies (straight-line, cost segmented depreciation), building costs, ordinance adoption, setbacks, myths, designs, sustainability, and useful policies/programs
• Served a secretary role in meetings and correspondence between stakeholders from the United States, Mexico, the UC San Diego Design Lab, and impact project leads to assist San Diego-Tijuana international collaboration projects
• Performed administrative duties for the Science & Tech Network, including event coordination, media design, monthly updates to the World Design Capital committee, event MC/moderator at events with community members & project leads

Real Estate + Data Science experiential learning fellowship with the Urban Studies & Planning Department and Homelessness Hub at UC San Diego, under the ULI H2H Mission Priority Initiative.

Affordable Housing Research Assistant

Mentors: Dr. Feiyang Sun, Dr. Mirle Rabinowitz-Bussell

• Reported affordable housing policy incentives and cost factors, emphasizing public improvements and tax incentives
• Networked with developers to study market feasibility, provided research team with empirical information on the barriers and opportunities facing multifamily projects, pertinent to timing, permitting, site contexts, parking, and construction, labor and materials cost.
• Received sponsorship to network with SDGBC to study green building techniques and integration with affordable housing

Data Science Fellow

• Programmed a bespoke algorithm to parse property data from CoStar into a tabular format for non-commercial purposes
• Applied statistics and machine learning on over 5000 LIHTC projects to derive factors influencing fund allocation
• Utilized analysis methods for social science: feature engineering, clustering, attribute stratification, marginal error analysis

Data science experiential learning fellowship with the Computer Science & Engineering department

Mentor: Dr. Niema Moshiri

• Developed an interactive application to visualize Covid-19 time-series data with variant segmentation
• Programmed a bespoke data-smoothing algorithm that selectively reads data upon user requests to optimize runtime
• Reduced initial runtime by 97% (5s to sub-0.1s) with caching/memoization and efficient data structure implementation
• Benchmarked data compression techniques to optimize serialization efficiency and expedite loading speed
• Wrote 7 research notebooks to uncover patterns in 30+ large-scale datasets involving 3000+ cancer patients
• Exceeded 87% accuracy across diverse evaluation metrics in predicting breast cancer recurrence status
• Analyzed disparities and data collection gaps to ensure ethical and responsible data science practices
• Honored with a nomination by Dr. Moshiri to deliver a presentation at UCSD’s 36th Annual Undergrad Research Conference
• Special commendation post by the Semiconductor Research Corporation on LinkedIn

Company website: Going Solo IT Services & IT Consulting

• Implemented a character-based word embeddings model (char2vec) based on a recurrent neural network using Keras from TensorFlow, coupled with Scikit-learn for linear dimensionality reduction of vectors
• Streamlined data labeling processes through the implementation of an automated Bayesian classifier, significantly improving efficiency
• Leveraged API calls to query data from findl.com, allowing seamless access to essential information for analysis
• Conducted comprehensive analysis of an S&P 500 company’s stock/financials and transformed insights into an interactive dashboard using Tableau, facilitating intuitive visualization and exploration of key metrics
• Demonstrated exceptional organizational skills by successfully coordinating projects and effectively managing team communication across 7+ time zones, ensuring smooth collaboration and timely delivery
• Enhanced project management practices by refining and maintaining a dynamic PM dashboard in Google Sheets, enabling real-time progress tracking and streamlined reporting to supervisors

Exceeded internship expectations by proactively implementing an automated data labeling tool by leveraging machine learning, effectively saving significant time and resources when the initial assigned approach was brute force, setting a new standard for efficiency and problem-solving at the company

Company website: Deloitte Business Consulting & Services

A data science mentorship program jointly organized by Deloitte and the Halicioglu Data Science Institute.

• Utilized ARIMA models to generate employment metric forecasts and predict market behaviors due to Covid-19
• Conducted in-depth analysis to extract actionable insights using multiple linear regression, exploratory data analysis (EDA), and visualization
• Collaborated closely with a dedicated mentor from Deloitte, receiving personalized guidance and support on a weekly basis
• Presented findings to an audience of Deloitte practitioners and advisors from the Halicioglu Data Science Institute

Assumed a leadership position within the project group, taking charge of task assignments, scheduling meetings, and cultivating a cohesive project story when our team needed leadership; fostering a positive, motivated team environment, on-time delivery, and quality assurance.

Program website: SPIS - Summer Program for Incoming Students

Pre-college summer computer science program

• Final Project: A web-app that generates vehicle suggestions given a user’s lifestyle following a decision tree
• Programmed in Python and HTML for brower display of user interface

Typed my first print(“Hello World”). Had zero clue what machine learning was, so I implemented a purely if-else decision tree to build a car recommender. Did manage to use Flask for front end display!

I volunteered to be my swim team’s manager (and swimmer) during my senior year in high school when we needed one

• Responsible for acquiring, storing, and presenting athletes’ race times and lap time splits
• Hardware facet: scoreboard operation, backup timers preparation, starter/speaker system setup
• Software facet: Excel data entry (styling must follow SCHS’s practices and in formats readable by sports software), Hy-Tek meet manager software operation
• Systems used: Colorado System 6 Sports Timing System and Hy-Tek meet manager

Notable accomplishments: 2018 JV Coach’s Award, 2019 JV Valuable Swimmer Award, 2021 Varsity Swim Appreciation Award, 2021 Varsity Coaches Choice Award

Fun fact: 4 awards given by 4 different coaches over 4 years