Jingwei He

San Jose, CA 95134 ยท jingweihe198@gmail.com

I'm a Computer Science master's student at Northeastern University with a strong background in mathematics and a strong interest in backend systems, data engineering, and scalable infrastructure. My recent work includes building a modular, cloud-based ETL pipeline orchestrated by Airflow and powered by AWS Glue (Spark) to efficiently process structured datasets in a fully serverless environment. I've also developed backend-focused web applications, which deepened my understanding of API design, system reliability, and cloud deployment. I'm currently seeking a 2026 Software Engineering internship where I can apply my skills to real-world systems and continue growing as an engineer.

Experience

Software Engineer Contractor

Xiamen Renyixing Network Technology Co., Ltd.

  • Built a retrieval-augmented generation (RAG)-based chatbot on WeCom for automated in-game chest inquiries and purchase assistance, reducing manual agent workload.
  • Designed the RAG database architecture with hierarchical information organization for GPT-4o, and implemented an HTTP webhook for real-time knowledge-base synchronization.
  • Integrated Dify with WeCom to enable multi-modal responses (text, images, and web content).
  • Optimized customer-support operations, cutting manual response time by over 80%.

    • Sep 2024 - Dec 2024

    Projects

    Cloud-based Airflow ETL Pipeline
    Python, Apache Airflow, AWS S3, AWS Glue, Spark, Docker

  • Built a modular ETL pipeline on AWS to automate extraction, transformation, and loading for large, evolving datasets, processing over 50 GB of raw data and eliminating manual workflows.
  • Used Apache Airflow to orchestrate AWS Glue (PySpark) jobs, automating data cleaning, timestamp parsing, and aggregation to generate partitioned Parquet outputs for efficient downstream analytics.
  • Developed a YAML-driven configuration system for flexible onboarding of new data sources, and integrated Protobuf validation to enforce schema consistency and prevent ingestion errors.
  • Created a UDF framework to enable users to inject custom transformation logic, supporting dynamic business requirements without code changes.
  • Designed to support over 5000 analytics queries in Athena and Redshift for business reporting.

    • Jan 2025 - May 2025

    Backend System for E-commerce Platform
    Java, Spring Boot, MyBatis, MySQL, Maven, Docker

  • Developed a backend system to streamline operations for a rapidly expanding e-commerce platform, designed to scale and support user authentication, product catalog, cart, order, and payment modules for over 10,000 users.
  • Structured data flow with MyBatis and layered architecture, reducing code coupling.
  • Improved security by implementing MD5+salt encryption and optimized performance with Guava caching and paginated queries, reducing response time by 40%.
  • Enabled dynamic category trees and product listings using recursion for fast search and flexible expansion..
  • Integrated Alipay for QR-based transactions and managed asynchronous confirmation flows, increasing payment success rate to over 99% and reducing transaction errors.

    • Aug 2024 - Dec 2024

    Full-Stack Social Media Platform
    TypeScript, Next.js, Tailwind CSS, MongoDB, Vercel

  • Developed a social media platform inspired by Threads to address real-time engagement and community features, which can support posting, threaded discussions, likes, follows, and personalized feeds for over 1,000 users.
  • Built a frontend with Next.js and TypeScript, delivering sub-second page loads and smooth user interactions..
  • Integrated Clerk for authentication, cutting implementation time by 40% and boosting login reliability
  • Utilized MongoDB to efficiently store and query, supporting high-concurrency access and scaling to 10,000+ posts.
  • Deployed on Vercel with automated CI/CD, reducing deployment and enabling global edge network delivery.

    • Dec 2023 - Mar 2024

    Instacart Product Recommendation
    Python, Scikit-learn, Pandas, ROC/AUC, F1-score, Cross-validation

  • Built machine learning models to predict recurring items in users' next order based on three million Instacart orders.
  • Preprocessed data by normalization, missing data value cleaning, quality checking, model label constructing.
  • Implemented logistic regression, gradient boosting, neural network, and user/item-based collaborative filtering.
  • Conducted hyperparameter tuning on gradient boosting models using cross-validation and the top 15 features.
  • Evaluated model performance via ROC-AUC (around 0.78) and F-1 score (about 0.38).

    • Aug 2023 - Dec 2023

    Education

    NorthEastern University

    Master
    Computer Science

    GPA: 4.0

    Jan 2025 - May 2027(Expected)

    University of South Florida

    Master
    Mathematics

    GPA: 3.66

    August 2020 - May 2022

    University of South Florida

    Bachelor
    Applied Mathematics

    GPA: 3.4

    August 2016 - May 2020

    Skills

    Programming Language
    Tools and Frameworks

    Publication