PhD Student at NYU Tandon

Most Relevant Technical Skills: Python, Pytorch, Sklearn, SQL, Machine Learning, and Deep Learning

Others: Web Development (Java, Vue3), C, C++, AWS, React, Django

Education

Ph.D., Urban Systems (Data Science and Machine Learning) New York University
M.S., Computer Science New York University
M.S., Accounting Fordham University
B.A., Math and Political Science (double major) Hamilton College

Publications

‘Human Intervention and Tree Health in NYC: A GMM Analysis’ (working paper)
‘Citi Bike Station Behavioral Regime Model and Its Application in Rebalancing Operations’(working paper)
‘NYC Evictions, Neighborhoods, and Gentrification’ (working paper)

Work Experience

Artificial Intelligence Intern | Wade Trim
– Enhanced the ChatBot project by integrating LLM APIs, RAG, and LangChain technologies to accurately respond to queries from selected PDFs and websites
– Designed new ChatBot interfaces using Figma and mapped out the backend workflow with LucidChart
– Collaborated with AI and SWE teams to brainstorm innovative and feasible AI products, actively participating in the coding and design processes using Jiva and Confluence

Student Data Engineer | New York University
– Imported university-wide survey data from database into python scripts, notebook, Tableau, and Excel
– Wrote Python Scripts to clean NYU Alumni employment datasets and analyze Alumni’s recent career trajectory

Software Engineering/Machine Learning Intern | SuperFun Technologies Ltd, Chengdu, China
– Employed Linear Regression, Decision Trees, Random Forest, and SVM techniques to analyze ICU patients’ post-surgical recovery data, enhancing understanding of recovery predictors.
– Partnered with an Emergency Room physician to analyze data and co-author a section of the hospital’s post-surgery patient care handbook, improving patient outcomes.
– Worked on a project parallel to the company’s actual multi-functional SaaS platform for a diverse team, optimizing food ordering, ingredient purchasing, and delivery processes. Technologies used include Spring Boot, Maven, Redis, MyBatis, Vue3, and Nacos.

Graduate Assistant | New York University
– Communicated with school administration regarding Tandon Graduate Admissions audit progress, assisted applicants navigate and troubleshoot applications

Projects

NYC Eviction Features and Trends from 83,282 records (current status: paper under review)

– Thoroughly cleaned and analyzed NYC eviction records, building attributes, 311 complaints, social vulnerability data, economic metrics, and gentrification proxies to examine their correlation with NYC eviction trends and patterns.
– Exhaustively investigated these 60+ potentially contributing factors through longitudinal (Time Series), spatial (ArcGIS), statistical, and machine learning techniques(Random Forest Regression, XGBoost, SHAP values, Linear Regression) at the unit, building, zipcode, and neighborhood level.
– Concluded that racial composition and socialeconomics were persistent and dominating factors, including historical redlining - Discovered that building types and maintenance were influential, such as buildings with 100 plus annual complaints were 40 times more likely to experince higher evictions; car ownership also appeared to be a previously unconsidered driver, as areas with higher car ownership (much higher expenses than using public transportation) correlated strongly with places with higher eviction rates.
– Indicated that tightened municipal requirements for rental unit maintenance, inspection, and fines could improve housing quality while decreasing evictions.

Generative Adversarial Network training on CelebA

– Created and evaluated a range of GAN models for generating realistic images, focused on diverse architectures like DCGAN, WGAN, and Relativistic GAN using the CelebA dataset in a team of 2.
– Conducted detailed experiments, including training variations and optimization techniques, to improve image generation quality and efficiency.

ResNet Models to train and test on CIFAR-10 image datasets for a Kaggle Competition

– Developed and evaluated various ResNet architectures for CIFAR-10 image classification in a team of 2, achieving a best accuracy of 97.0% on test data.
– Conducted extensive experiments on hyperparameters and model variations, documenting results and optimization strategies in Jupyter Notebooks.

High School Finder App “Shortlist” – Python, Vue, Django

– Contributed to the development of Shortlist™, a web application that matches middle school students and their parents with high schools in NYC, using Vue, Django, and AWS technologies.
– Collaborated on diverse aspects of the project including UI/UX design, backend integration, and deployment, enhancing the application’s functionality and user experience.

Airbnb House Price Prediction Model

– Developed a predictive model for Airbnb rental prices using machine learning techniques in Jupyter Notebook.
– Analyzed and visualized housing data to identify key price determinants, enhancing accuracy in price prediction models.

Java Yahtzee DB Socket Project

– Developed a Java-based Yahtzee game with database and socket integration for real-time multiplayer functionality.
– Implemented robust game logic and GUI, and managed backend operations for storing player statistics and game scores.

Airbnb DataBase SQL Python Streamlit Project

– Developed a Q&A model for Airbnb rental prices using SQL and Python.
– Analyzed and visualized housing data using StreamLit frontend functions help customers to identify critical pricing and neighborhood information to better prepare trips.

About Me ✊

Hobbies

– Reading
– Photography
– Road Tripping
– Badminton or any other sports with team spirits
– Running: Brooklyn Beach Half Marathon, 2019; Skunk Cabbage Classic Half Marathon, 2018; Syracuse Empire Half Marathon 2017