Data Science Analytics

XIAO DENG

Los Angeles, CA ***- ***- **** ********@*****.***;*****@***.*** https://www.linkedin.com/in/xiaodeng1/ SUMMARY

Self-driven USC student studying Data Science and Economics who demonstrates proficiency in programming and statistical analysis. I bring practical experience with Python, SQL, Power BI and Excel while having worked on projects related to customer behavior analytics, smart support systems using NLP, and real estate investment analysis. Proven capacity for self-directed learning of complex technical skills through independent completion of AP-level calculus and statistics during high school. I am enthusiastic about implementing data-driven solutions that solve business problems and guide strategic decisions. TECHNICAL SKILLS

Programming: Data Cleaning & Wrangling (Python), ETL (SQL, Python), NumPy, pandas Databases & Tools:MONGO DB,Mysql, SQL

Visualization Tools: Tableau, PowerBi, Python: Matplotlib Modeling: Supervised Learning Methods (Linear & Logistic Regression, Decision Trees-CART, CHAID, Random Forest, SVM) Tools:Microsoft Excel, Microsoft office, Photoshop PR Others:Azure, Shopify, Google Analytics, GIT, AI Model Localization; AI Server Establishment,Gradio, RAG Language: English, Mandarin, Cantonese.

EDUCATION

University of Southern California Los Angeles CA

Bachelor of Science Economics and Data Science (STEM) 3.52/4 Aug 2024-Dec 2026 Minor:AI Applications

Coursework: Accelerated Programming in Python, Data Analytics University of California Santa Barbara Santa Barbara CA Bachelor of Science GPA 3.97/4 Double Major in Economics And Data Science(STEM) August 2023-June 2024 Coursework:Advanced Algebra,Vector Calculus,Probability and Statistics, Data Science, Principles of Microeconomics,Microeconomic Theory

WORKING EXPERIENCE

Tianhe Digital Mall Guangzhou, China IT Engineer(intern) June 2024-August 2024

Built an Excel-based customer info system with VLOOKUP/macros, improving service efficiency by 80%.

Created an internal knowledge base for fault diagnosis, improving troubleshooting accuracy to 90%.

Installing computers for clients and delivering after-sales services, including hardware integration, software installation, network configuration, fault diagnosis, and resolution The Afiliated High School of The SCNU(Guangzhou, China) Teacher's Assistant Part Time August 2022-June 2023

Supported classroom operations, including attendance, grading, and individual tutoring

Assisted in lesson preparation and identifying student learning challenges

Built a grade analysis system using MySQL, Excel, and Word data; cleaned data with Python (Pandas) and visualized insights in Tableau to generate personalized learning reports

Volunteer Traffic Coordinator Guangzhou, China 2022–2023 Guided pedestrians and ensured road safety compliance. Provided location consultation and assisted elderly and disabled pedestrians. PROJECT HIGHLIGHTS

Built a customized Shopify store with 3D preview, Google Docs integration, and behavioral analytics.

Developed a local AI chatbot using Qwen and RAG to automate domain-specific customer support.

Modeled IRR/NPV for real estate investments using financial statements and risk assessments.

Created a virtual goods trading system using Python and APIs, achieving 40% annualized returns. REWARDS & HONORS

AMC(American Mathematics Competitions) Highly distinction(top1%) BPhO(British Physics Olympiad) Bronze award

AP Scholar with Distinction: Self-studied and passed 9 AP exams PROJECT EXPERIENCE

Real Estate Investment Financial Modeling & Risk Assessment Build a financial model based on financial statements and supplemental information. including lRR and NPV, to identify the risks and benefits and decide whether a project-level real estate equity program is worthwhile Key Responsibilities:

Read through the balance sheet,income statement, and cash-flow statement and evaluate the company's situation based on some key values, e.g., gross margin, interest expense, debt-to-asset ratio, and operating cash flow.

Evaluate the risk based on the supplemental information such as debt schedule, pre-sale report, etc.

Build a financial model based on all the information and calculate the NPV and IRR.

Help making decisions based on data.

Quantitative Virtual Goods Trading:

Making data-driven trades based on a price forecasting financial model(SMA) and pricing information obtained by programming languages integrating APls from websites and made a 40% CAGR(compound annual growth rate). Key Responsibilities:

Use Python to integrate APls and grab pricing data from different platforms.

Extract and Transform data into a dataframe and find the optimized buying price.

Build an SMA model to predict the price for each good based on past trading information and make trading decisions based on the modeled return rate.

Automatic updates and visualization.

Shopify Custom Store Development Project

The Shopify storefront underwent customization that enabled interactive product displays alongside time-based consulting services complemented by client-specific document sharing capabilities. Key Responsibilities:

Developed a functional online store on Shopify for internal testing and client demonstrations

Created custom Metafields to enable 3D product previews and integrate shared Google Docs resources

Wrote embedded code to implement advanced store functionalities beyond Shopify’s default capabilities

Integrated Google Analytics and Microsoft Clarity for real-time user behavior and conversion tracking

Customized GA reports and dashboards to meet individual client KPIs and analytical needs AI-Powered Customer Support Chatbot (Qwen2.5:1.5b + RAG + Gradio) Developed a local LLM-based intelligent customer support system to reduce workload for human agents and provide 24/7 high- quality service, while ensuring data privacy and full on-premise control. The system is planned to be integrated with the company’s e-commerce platform for seamless automated support across order, logistics, and refund modules. Key Responsibilities:

Designed and deployed a local AI-powered customer support system using LLM (DeepSeek/Qwen) with a Gradio web interface for multi-turn dialogue.

Implemented a RAG (Retrieval-Augmented Generation) pipeline by collecting and preprocessing historical customer service data, and indexing it using SentenceTransformers and FAISS.

Led the collection, cleaning, and structuring of domain-specific Q&A datasets to improve model relevance and retrieval accuracy.

Evaluated response quality through manual testing and prompt-tuning, iterating on data preparation to enhance performance in real-world customer service scenarios.

Contact this candidate