Basic Information

Name: Ruihan Xia
Date of Birth: September 13, 1993
Region: Chongqing, China
Work Experience: 6 years
Job: Internet R&D engineer
Portfolio Website: www.xiaruihan.xyz

Education

Bachelor's: Southwest Petroleum University - Petroleum Engineering June 2012 - December 2016 (GPA: 3.4)
Masters: Beijing University of Aeronautics and Astronautics - MBA June 2020 - December 2022 (GPA: 3.66)

Publications

Undergraduate Thesis - Application Research on High-Density Cement Slurry to Prevent Gas Kick
Master's Thesis - Machining learning stock selection research under the scenario of industry rotation

Work and Project Experience

2015.12 - 2018.01 | JD.com | Store BI Backend Developer
JD.com is a leading e-commerce platform in China. The JD convenience store BI project aims to provide data services for offline stores integrated with JD. In this project, I mainly played a role as a backend developer, performing the following tasks:

• Data Warehouse Construction: Using HiveSQL, I constructed the fact layer and dimension layer involving tables for orders, inventory, stores, and staff. Additionally, an aggregation layer was established to optimize data query performance. Using Sqoop, data was regularly synchronized to MySQL, ensuring real-time accuracy.
• Permission System & Retail Store BI Backend Development: Leveraged Spring Boot and MyBatisPlus to set up the project framework, simplifying database operations. Integrated with Spring Security, fine-grained permission management and core business logic were realized to meet customized requirements. To enhance query speed, a Redis caching strategy was implemented to optimize frequently queried data.

2018.01 - 2021.05 | Mioji Travel | Strategy Development
Mioji Travel utilizes AI to plan global travel routes. I was responsible for formulating and developing strategies. Key strategies include:

Travel Route Pricing Strategy for Mioji:

• Background: global routes comprise air tickets, hotels, and leisure activities. Each product has multiple supplier prices. It's necessary to determine the actual cost for each product, the reference price for the clients, and then set a quotation for them.
• Strategic Thought: Price for customer satisfaction and profitability for Mioji.
• Strategy Formation: Used hierarchical clustering to determine reliable cost prices. Based on market sensitivity, established a reference price that aligns with client expectations. Defined profit margins and maximized a satisfaction function, using hill-climbing algorithms to finalize the product's pricing.
• Strategy Implementation: Used Scipy's hierarchical clustering for reliable quotes. Implemented satisfaction function and hill-climbing algorithm in Python to get product pricing. Also, set up an API using Flask for business-end access.

Itinerary Creation Strategy

• Background: After the AI designs global travel routes, there's a need for textual descriptions. We aimed to design a system that auto-generates these descriptions.
• Strategy Idea: Break down elements of the journey, consider specific combinations of elements, and use preset phrases to form articles.
• Strategy Development: Using Python, we recognized elements and patterns between them. For identified patterns, we randomly selected phrases to create articles, which were then provided to the business side.

2021.05 - present | ZuoyeBang | User Profiling Backend & Tag Development
ZuoyeBang is China's largest K12 education company. I participated in the backend development of user profiling and the design of user tags.

• Function Development: This includes functions like audience selection, audience description, and tag management. I used Spring Boot to build the server side, loaded Hive tag tables into ES, and queried user information from ES. User groups were then generated and stored in Tencent Cloud COS for other business use.
• Tag Development: I developed basic tags using HiveSQL, such as basic attributes and value attributes. I also developed complex tags based on Xgboost to predict student course purchase intentions. Using 70+ user tags as features and recent course purchase data as labels, I trained a classification model, using positive sample probability as the output. The tag's effectiveness was determined through business AB testing.

Independent Side Projects

2023 | AI Paper Management System

• A program that simplifies core functionalities of Zotero and Citespace.
• The program can download paper metadata and the full text by titles. It can cluster topics using a community algorithm and identify key points through intermediacy centrality. Lastly, it can utilize AI to analyze papers in bulk.

2021 | NuanYiYi Psychological Counseling

• As an independent developer, I designed and developed a WeChat mini-program for psychotherapists.
• Search "暖一一" to view the Wechat mini-program. This project won first place in the Beihang University Entrepreneurship Competition.

2023 | AutoClaude Simplified Version

• A more interpretable autogpt.
• This reduces the complexity of AutoGPT, incorporates the Claude 100k API, and combines it with Google AI to automatically fetch information.

Honors

• Southwest Petroleum University - Outstanding Graduate (2016)
• JD.com (Jingdong) - Future Star Excellence Award (2017)
• Mioji Travel - Best Model Employee (2019)
• Zuoyebang (Homework Help) - "Enterprising" Values Award (2020)
• Beijing University of Aeronautics and Astronautics - First Prize in Campus Entrepreneurship Competition (2021)
• Tableau Certified Data Analyst (2022)