Andy Cui Profile Pic

Andy Cui

Name:
Andy (Xiang-Yu) Cui
Current Address:
Waltham Massachusetts
Phone:
(402)-853-3000
Email:
xiangyucui@outlook.com

About Me

Senior Full Stack Software Engineer King 7 Club Corp

Integrating state-of-the-art design and development practices, I specialize in creating robust Machine Learning, back-end, and front-end applications, having delivered multifaceted projects across diverse domains. My technical toolkit is comprehensive, encompassing Java, Python, MySQL, PostgreSQL, AWS, React.js, and adept implementation of CI/CD pipelines. Mastering complex problem-solving, I am a self-driven technophile, relentlessly chasing the latest technological breakthroughs. As a fervent learner in Machine Learning and Natural Language Processing, I am particularly captivated by the prospects of mining cutting-edge big data and transforming it into insightful visual representations.

AWSAWS
DockerDocker
SupabaseSupabase
CursorCursor
DeepseekDeepseek
GPT2GPT2
PythonPython
NodeJSNodeJS
JavaJava
UbuntuUbuntu
GentooGentoo
MySQLMySQL

Experience

King 7 Club Corp     

Senior Full Stack Software Engineer - Full Time       Jan 2025 - Present, Los Angeles, CA, United States

  • Developed and deployed a responsive website using Node.js for the frontend and FastAPI for the backend, with component-based architecture for modular scalability. Implemented dynamic UI with custom JavaScript logic and CSS animations to ensure a seamless user experience across devices and languages.
  • Hosted static assets via GitHub, containerized the full-stack environment with Docker, and deployed the application on AWS EC2, ensuring consistent and reproducible builds across development and production. Utilized PostgreSQL as the backend database to securely manage and store user data.
  • To optimize access for both international and Chinese users, implemented intelligent DNS-based traffic routing: international traffic is served through AWS Global Accelerator, while Chinese users are redirected to a mirror deployment on Alibaba Cloud. This global traffic segregation strategy reduced cross-region latency by up to 90%, significantly improving page load times and user experience across all target markets.
  • Integrated Google Analytics Reporting API to track traffic and user behavior, enabling the development of interactive dashboards to monitor key metrics such as page views and user engagement across platforms like TikTok, Red Book (Xiaohongshu), and YouTube. This real-time data integration streamlined reporting workflows and automated insight generation, improving backend operational efficiency by 80%.
  • Enhanced backend performance and frontend delivery using CDN and caching strategies, and structured API responses in JSON format for clean integration.
  • CAC Auto Group LLC     

    Software Engineer - Full Time       Feb 2024 - Dec 2024, Southborough, MA, United States

  • Developed and maintained a predictive pricing system for vehicles on CarGurus using AWS serverless architecture, enhancing market compatibility and streamlining operations. Leveraged key AWS services including S3, Lambda, DynamoDB, SNS, CloudWatch, and Kinesis, and used Python with AWS CloudFormation for scalable infrastructure deployment.
  • Designed and implemented a fully serverless data pipeline to continuously monitor target data sources using Kinesis streams and Lambda triggers, eliminating the need for traditional polling. This approach reduced infrastructure and processing costs by 80%, while maintaining high scalability and responsiveness.
  • Integrated real-time monitoring to track market data fluctuations, enabling automated detection and adjustment of vehicle prices in response to deviations. This solution boosted daily operational efficiency by 80% and improved pricing accuracy by over 50% compared to industry standards.
  • AlpaLifeBio LLC     

    Software Engineer of Data Engineer - Intership       Dec 2022 - Jun 2023, Woburn, MA, United States

  • Built and managed a robust AWS streaming data pipeline to automate biomedical data ingestion from multiple public databases into Kinesis Data Stream. This system processed over 500,000 data entries daily, using Lambda Functions for real-time data transfer and S3 and DynamoDB for efficient, scalable storage and retrieval. This architecture allowed seamless handling of high-volume data with minimized latency and reduced operational costs.
  • Configured and optimized a structured SQL database to integrate and process data from diverse biomedical sources. Implemented an efficient tag-processing system for enhanced search and retrieval operations, reducing data retrieval time by 80%. This improvement significantly boosted operational efficiency, making it easier to access and analyze critical information for downstream applications.
  • Applied advanced data matching algorithms and TF-IDF (Term Frequency-Inverse Document Frequency) techniques to perform data comparison and accurately identify potential client profiles with a 95% match rate. This methodology directly supported targeted marketing strategies by enabling precise identification of high-value clients within the biomedical field.
  • Enabled real-time notifications for data updates by setting up DynamoDB Streams to capture table modifications. Configured Lambda Functions to process these events and send automated alert emails through Amazon SES, enhancing response times by 30% and streamlining the data update workflow.
  • Dutchgo LLC     

    Software Engineer - Self-employed & Co-founder       Oct 2020 - Apr 2022, Omaha, NE, United States

  • Developed a customer demand analysis pipeline during the Covid-19 pandemic using AWS S3 for secure data storage and Lambda for serverless processing, combined with SQL and Python. Built and deployed a Random Forest model on AWS SageMaker to identify high-potential user profiles for conversion, achieving an accuracy of 98%. This automated approach saved considerable processing time and improved prediction accuracy for strategic marketing efforts.
  • Established a data ingestion and processing framework using AWS Glue to automate data ETL processes, enabling seamless integration of monthly customer market data from S3 into DynamoDB for structured storage. Employed SQL and Python to preprocess data and built a multiple regression model on SageMaker to predict revenue trends, achieving an 85% accuracy rate. This solution provided valuable insights for strategic decision-making regarding resource allocation and customer segmentation.
  • Analyzed user behavior patterns to inform UX design by employing Human-Computer Interaction (HCI) principles. Utilized Axure RP to create interactive software prototypes aligned with user behavior insights. Additionally, used AWS Amplify to deploy and test front-end prototypes rapidly, ensuring seamless integration with back-end data and enhancing user experience.
  • Constructed real-time monitoring dashboards using Tableau to track essential investment metrics, such as ROI (Return on Investment), growth rate, and portfolio balance. Integrated AWS Kinesis for real-time data streaming, enabling prompt anomaly detection and root cause analysis. This system provided stakeholders with actionable insights on data trends, supporting a 30% faster decision-making process for adjusting business strategies.
  • Set up AWS CloudFormation templates to manage and automate infrastructure provisioning, ensuring efficient resource management and scalability of the analysis system. Leveraged AWS IAM to implement robust access control, securing data and model resources across the team.
  • Chatchup(Zhuiguang) Information Technology Co. Ltd     

    Project Manager & Project Analyst       Jan 2019 - Sept 2020, Henan, Zhengzhou, China

  • Analyzed customer requirements to draft an activity diagram and outlined preliminary UML page functionality scenarios. Initiated the development of customized software (Web & App) tailored to customer needs.
  • Employed advanced SQL queries and Python to analyze over 50GB of market data on similar products, identifying customer usage patterns and business requirements.
  • Conducted exploratory data analysis on numerical, categorical, and time-series data using Matplotlib and Seaborn. Constructed multivariance Time Series Clustering model, DTW (Dynamic Time Warping) with Hierarchical clustering, on 120 features to group customers, segment market and described the clustering centroid by DBA (DTW Barycenter Averaging).
  • Trained market demand models, including Isolation Forest and Robust PCA (Principal Component Analysis), as well as Linear Models and Random Forests, to establish a system for identifying and responding to customers' potential requirements.
  • Constructed dashboards with Tableau to depict customer usage activities and submitted analytical reports to clients.
  • Dell Inc     

    IT Support Specialist       Oct 2013 - Oct 2015, Dalian, LiaoNing, China

  • Organized on-campus after sales technology support events for Dell products.
  • Improved reports procedures and saved five hours of manual work every week.
  • Managed an average of ten requests daily, creating and updating data elements on the Dell Campus platform.
  • Education

    Northeastern University     

    Master of Science in Artificial Intelligence Dec 2023

    Relevant Coursework:

    Object-Oriented Design, Natural Language Processing, Game Artificial Intelligence, Machine Learning, Computer/Human Interaction, Algorithm, Wireless Sensor/Internet Things, Deterministic Ops Reaserch, Database Managemen

    University of Nebraska Lincoln     

    Bachelor of Computer Science May 2020

    Relevant Coursework:

    Wireless Communication Networks, Automata Theory, Unix Programming, Discrete Structure, Internet Systems & Programming, VR Programming of Unity, Numerical Analysis, Data Structure & Algorithm, Operating Systems Kernels, Basic Programming

    Awards: Authentication Place in ICPC ACM 2017 Central American Finals

    Dalian Neusoft University of Information     

    Eletronic Infomation Engineering July 2015 - Transfered to University of Nebraska Lincoln

    Relevant Coursework:

    Embedded System Programming, Sensor & Signal Theory, C Basic Programming, Analog Circuit, Digital Circuit, CAD Engineering Drawing, PCB Design with Auto CAD, TCP/IP Network Programming

    Awards: 1st Prize in Computer Business Entrepreneurship Contest, Winner of 2015 Freescale Smart Car Competition

    Skills

    Java85%
    Python90%
    JavaScript80%
    HTML85%
    CSS75%
    C/C++70%
    Assembly60%
    VB65%
    MySQL90%
    PostgreSQL85%
    MongoDB80%
    Redis75%
    SQLite80%
    TF-IDF85%
    Naive Bayes80%
    GPT-275%
    Bert80%
    CNN85%
    Transform80%
    Spring Boot85%
    React.js80%
    Node.js85%
    jQuery60%
    GitHub60%
    TFS60%
    Azure DevOps60%
    Jenkins60%
    Jira60%
    AWS60%
    Docker60%
    Maven60%
    Tomcat60%
    Axure60%
    Servlet60%
    JUnit60%
    Nginx60%
    Office60%

    Projects

    Employment Website Design

    Employment Website Design

    – A Web-based Axure based on CRAP interface interaction

    Designed a co-op-focused Employment Website using HCI principles to offer an intuitive experience catering to non-technical users, ensuring broad usability.
    Applied CRAP design principles to create a visually cohesive and navigable interface, addressing existing website pain points by incorporating critical features such as visa type and field of work.
    Led the unification of UI design across the platform, ensuring a smooth and accessible user experience for a broad audience. Validated design success and user satisfaction with thorough usability testing, confirming high effectiveness and gratification.
    View Project
    Amazon QA robot trained by BERT & GPT-2

    Amazon QA robot trained by BERT & GPT-2

    – Trained amazon toy customers' comment to train a QA robot

    Developed a Question Answering (QA) system using GPT-2 and BERT models focused on Amazon product reviews, assessing their performance in natural language processing tasks.
    Applied advanced NLP techniques, including TF-IDF Vectorization, CBOW, and Skip-gram, for effective data preprocessing and analysis.
    Conducted K-Fold Cross-Validation and ablation studies to optimize model performance and robustness, utilizing comprehensive model training and fine-tuning with a Kaggle dataset. We focused on accuracy, LOSS, and BLEU score metrics for evaluation, ultimately achieving high precision with an Accuracy F1 score of up to 94%.
    View Project
    Online Shop Database

    Online Shop Database

    – A online shop database system, deploy by SQL and python.

    Developed a Python and MySQL-based Robot Store Management System to streamline operations for small supermarkets, featured in a terminal user interface.
    Implemented SQL database structures and Python-MySQL connections, enabling efficient user account, order, and refund management.
    Constructed and documented a secure, offline management system, accommodating customer and operator interactions with potential for web expansion using Django.
    Managed project version control with Anaconda and maintained code repository on GitHub, with comprehensive end-user documentation and test cases.
    View Project
    Crash Loyal AI Agent collision method

    Crash Loyal AI Agent collision method

    – An AI agent's collision method implemented Crash Loyalty game

    Implemented character behavior logic that prioritizes enemy engagements and tactical decisions, ensuring actions adhere to game rationality and strategy.
    Conducted thorough testing and debugging to maintain game balance and performance, ensuring characters operate within defined routes and adhere to game boundaries.
    Orchestrated a dynamic unit deployment and resource management mechanic, incorporating a strategic elixir accumulation system for gameplay depth.
    Managed project documentation and version control, ensuring robust testing and functionality within the set development timeframe.
    View Project
    Doctor Lucky java Jframe game

    Doctor Lucky java Jframe game

    – An OOD use MVC implemented by java Jframe game

    Implemented character behavior logic that prioritizes enemy engagements and tactical decisions, ensuring actions adhere to game rationality and strategy.
    Conducted thorough testing and debugging to maintain game balance and performance, ensuring characters operate within defined routes and adhere to game boundaries.
    Orchestrated a dynamic unit deployment and resource management mechanic, incorporating a strategic elixir accumulation system for gameplay depth.
    Managed project documentation and version control, ensuring robust testing and functionality within the set development timeframe.
    View Project
    Gobang AI Game Agent

    Gobang AI Game Agent

    – An alpha-beta AI Agent implemented with Bobang

    Implemented a Gobang AI agent leveraging alpha-beta pruning to calculate stepwise rewards, enabling the agent to strategically discard less rewarding moves for enhanced gameplay efficacy.
    Architected predefined Gobang chess strategies with assigned value metrics, which were intricately incorporated into the AI's decision-making process, augmenting the competitiveness of the agent.
    Synthesized traditional Gobang tactics with algorithmic precision, optimizing the AI's play style to reflect a deep understanding of the game's cultural strategy elements.
    View Project