QQ音乐推荐算法实习生

• Team Diagnostics Analysis of Synthetic Training: Built an end-to-end machine learning model pipeline, from data extracting, preprocessing to ultimate modeling. Classified, generalized the pipeline to diagnosed team performance with high accuracy and small synthetic or real-time training samples (e.g. for army training).• Designed the activity tree to aggregate individual activities (STEP) as features to the top (SESSION, team result)• Tech Stacks: xAPI Activity Tree, MongoDB, OOD, Learning Records Store, Semi-Supervised Learning

• Student Engagement and Teacher Productivity App: Aimed at K-12 education by delivering a 360-degree engagement E-Learning platform among students, teachers, parents, publishers, and administrators• As backend team leader, designed the data model (ER Diagram with 18 entities and 4 roles) with CTO from scratch. Helped schedule and assign the backend development tasks. Documented the API Swagger notes for front-end team• As backend developer, developed complex CRUD services, resources (Swagger APIs) and DAO layers (JDBC Hibernate query) using Jhipster with Spring Boot framework, enable to run most of features and API of prototype• Designed DTOs, and used JWT and PreAuthorize methods to solve security issues• Tech Stacks: JHipster, Spring Boot, API Design, Data Modeling, H2, Postgres, Elastic Search, CICD, Git

Intern, Network Management Center, Data Modeling Team• Fraud Calls Detection: Extracted over 230K unbalanced customer profile data from Hadoop using Hive. Built and improved classification model (LR) performance using feature engineering techniques (Information Value, etc.) • Model Deployment: Provided RESTful API and deployed the Spark classification model on Company’s pipeline

Intern in the research project team at Digital Bank Innovation Lab.Project: Text Generation of Fintech News Summary• Automated the latest Fintech news fetching process with a crawler in Scrapy, and Splash and Selenium frameworks from multiple social media websites. Filtered, de-duplicated and ranked, crawled data via text mining and Natural Language Processing algorithms (e.g. AP clustering, TF-IDF, Text Similarity, etc.)• Utilized TextRank method to extract keywords and generate recommended news abstract

Intern in Oracle IT consulting team.Project: Development of ERP Analysis Platform of XIANG YU Group• Data Analysis: Analyzed data from 6 financial analysis BI tables based on financial reclassification and requirement documents logic. Updated test reports weekly• Data Extracting & Aggregating: Tested data accuracy through several complete ETL pipelines by using SQL.

• Edited Excel manuscript of semi-annual financial statement .Conducted computation & analysis of financial index.• Conducted financial statement analysis and industry research of banking industry. Analyzed why and how revenuestructure changed& figured out the key profit drivers. Structured semi-annual reports in consulting way.• Utilized PCA method with R and SPSS to evaluate the scientific accuracy of performance model & managed the data, explained the principle behind this model to staff, extracted key factors that affected factor score.