Chunyi G. Email & Phone Number

San Jose, Ca, Us

美国 加利福尼亚州 圣荷西

San Jose

• Built data flow sdk for data streaming for a large volume of video; focused on data processing, binary serialization and deserialization using Kafka, Faust and Msgpack• Wrote an object-detection application on top of flow-sdk with yolov3, in 3 components: video reading, image processing, image rendering with bounding boxes. Preprocessed data for same resolution due to buffer size limit by ffmpeg for input data coding and output data recovery• Developed a movie-recommendation application, to validate the feasibility of multiple data types, set incoming user info as stream data, movie and user’s history as batch data, processed them and output recommendations, and wrote test script using Pytest, sent feedback to all team members• Implemented Docker image to integrate flow-sdk on platform and used helm to deploy applications on Kubernetes system• Wrote some RESTful APIs using Flask, auto built and deployed micro services on server, check services’ status• Visualized deep learning model by transferred frozen pb file using python and Tensorflow platform, displayed model graph and analyzed operations, added two test cases to verify calculation’s correction and especially for customers: random cases generated by Numpy package, targeted cases based on customers’ model• Replaced fully-connected layer with conv2d operation which can help the design light-weighted and multi-purposed• Generated random and valid 64bit hex number representing different isa instructions by python program and pandas, argparse lib; transferred csv structure to dictionary for future use, saved 10 engineers’ hours every week in the team• Added generator and diff scripts into test flow, checked outputs from hardware and software simulation, translated data to readable JSON file, sent inconsistent results to designers, increased testing efficiency• Built CI/CD flow with YAML config file under a docker environment output in Gitlab

中国 江苏 南京

-Responsible for Installation and configuration of Hive, Sqoop, and on the Hortonworks Hadoop cluster using Ambari-Involved in extracting large sets of structured, semi-structured and data from various sources like FTP servers, API calls-Developed and implemented Hive scripts for transformations such as evaluation, filtering, and aggregation-Developed Sqoop scripts to import data from the database and created Hive external tables and views, on the data imported into the HDFS. Load real-time data from various data sources into HDFS using Kafka

中国 江苏 南京

-Defined users’ requirements and built web application models with customers based on their needs-Developed templates to generate and classify customers’ comments, ratings, transferred them into graphs-Implemented RESTful server-side endpoints user interface and database services with modern web frameworks-Integrated REST APIs with front-end Pages by generating JSON data from MongoDB database-Involved in debugging HTML5, CSS3 and JS code using tools such as Chrome Inspector-Maintained company website with PHP, HTML, CSS, and Bootstrap.

Master'S Degree, Computer Science, 4.0/4.0

Bachelor'S Degree, Network Engineering, 3.45/4.0

Activities and Societies: 学生会外联部部长