Tina Bu

The Kane Regional Center is a nursing and rehabilitation care provider. I worked as a student IT consultant and digitalized the donating process for their foundation, and the working logs for the security department.

• Evaluated the architecture of a security logbook software under development in terms of quality attributes, functional requirements, technical and business constraints, designed a database schema and outlined the developing roadmap for a more decoupled and reliable Java GUI application with MySQL as the database
• Implemented a CRM system for Kane foundation and digitalized the whole donation and event organizing process

JD.com is the biggest B2C company in China, also the biggest competitor of Alibaba.

• Performed daily website traffic analysis of JD's overseas market on Apache Hive platform to customize client interactions
• Designed metrics and analyzed sales campaigns to guide product page layout, coupon, marketing and ad placements
• Contributed to the collaborative filtering recommending system item-based module which promotes total sale by 11%
• Tracked and blacklisted more than 1000 fraudulent accounts and restricted coupon usage and sales products purchase

Teradata provides data warehouse solutions as well as analytics platforms

• Configured Hadoop clusters; supported the proof-of-concept demo for a potential customer through performing ETL and data cleansing, developed in-database MapReduce functions, and employed the SQL-MR’s “nPath” for graph analysis
• Supported field application engineers on big data solutions for high profile projects including churn prediction and precision marketing for a telecom company, and public loan risk analytics for the Ever-Bright Bank

Coursework: Machine Learning, Machine Learning with Large Datasets, Big Data Analytics, Practical Data Science, Research in Practical Data Science, Software Architecture, Data Intensive Scalable Systems, Database Management, Data Structure, Information Security, IT Consulting

Outstanding Leadership Award | Academic Scholarship | GPA: 3.7

* Designed a threat detection and access control RESTful API in LISP using a knowledge-base AI system Scone

* Led a team of 4, actively scheduled the project to view the risks and mitigate them; orchestrated for the communication and synchronization between mentors, advisors, and customers to gather requirements and update project status; guided and documented the project through 3 system architecture design iterations, and open-sourced it to the community

* Developed an e-commerce RESTful web API application that processes over 4 million records and returns a keyword search result with median 4.4 ms using Node.js, MySQL, MongoDB, and ElasticSearch; bench-tested with Artillery

* Deployed the server on AWS with Elastic Load Balancing (ELB) and AutoScaling Groups, RDS with subnet groups and ElasticCache and achieved an average performance of 10,000 RPS with distributed transactions and concurrency

* Extracted deep learning features from Pittsburgh satellite images with the pre-trained VGG-16 network and classified the images by wealth level using CNN achieved 79.7% accuracy with the baseline softmax reg accuracy being 24%

* Proposed as an alternative wealth and poverty census data collection method with significant cost and time savings

* Trained an SVM classifier for geographical sentiment analysis of the 2016 presidential candidates Clinton and Trump using tweets extracted randomly from different states across the U.S., achieved 85% accuracy on the cross validation set