Songxiang Gu

Director at JDT
  • Claim this Profile
Contact Information
us****@****om
(386) 825-5501
Location
Boston, Massachusetts, United States, US

Topline Score

Topline score feature will be out soon.

Bio

Generated by
Topline AI

5.0

/5.0
/ Based on 1 ratings
  • (1)
  • (0)
  • (0)
  • (0)
  • (0)

Filter reviews by:

John Cardente

During the summer of 2008 Songxiang Gu completed an internship with the Celerra Division of EMC Corporation. Specifically, Songxiang interned with the NAS Advanced Development group, a high performance engineering team chartered with developing new technologies and innovations for EMC's network storage product line. Throughout the internship I acted as Songxiang's manager responsible for defining his internship project, assigning tasks, and monitoring progress. For his internship Songxiang was given the task of creating a suite of software utilities for analyzing network traces of distributed file system protocols, specifically NFS and CIFS. Completing this project required Songxiang to quickly become familiar with both protocols, determine their semantic similarities and differences, and develop a method for representing both protocols in a common structured XML format. Songxiang then developed software utilities for converting protocol specific NFS and CIFS traces into the common format he defined. In addition, Songxiang wrote a variety of software utilities to analyze common format files including file system simulators, activity profilers, and statistical visualization tools. In total, Songxiang implemented 15K lines of C++ source code which, when compiled, produced a suite of Linux user-space command-line applications. Songxiang provided ample documentation for his software and utilized a consistent testing methodology to ensure its quality. During this work, Songxiang demonstrated significant engineering and software development skills. Throughout the internship Songxiang acted with a degree of professionalism far beyond that expected of an intern. He was organized, highly productive, and proactively provided weekly reports detailing his progress and plans for the following week. On a personal level Songxiang presented a friendly manner and interacted frequently with other group members in various social settings. By the end of his internship Songxiang was a fully integrated member of the team and a welcome presence. In summary, Songxiang successfully completed his internship despite many challenging assignments. He was professional, productive, and pleasurable to work with. He demonstrated a significant level of skill and an ability to quickly learn complicated topics. In summary, Songxiang is an excellent software engineer with valuable skills that I would not hesitate to hire again for the same or similar position.

You need to have a working account to view this content.
You need to have a working account to view this content.

Experience

    • China
    • IT Services and IT Consulting
    • 300 - 400 Employee
    • Director
      • Jan 2021 - May 2022

      • Led a team to develop a home-made federated learning system, make contributions in architecture design and team management, finally open-source the project at https://github.com/fedlearnAI/fedlearn • Published papers in the federated-learning field in KDD

    • Senior Software Engineer Manager
      • Aug 2019 - Dec 2020

      • Led a team to develop a Knowledge Graph System for Fin-Knowledge Analysis, including the models of NL2SQL, Triplet Extraction, TransX Embedding and Causality Analysis. • Led NLP team to push the automatic customer service system to next level by adding multi-run conversations, chatbot configuration, model hot-deployment, A/B testing as well as connecting to phone system through ASR/TTS.

    • Team Lead Manager
      • 2019 - Jul 2019

      • Led NLP team to develop an auto-customer-service platform for both customer service and chatbot-sales purpose to boost our business. • Developed an AI online serving platform with distributed training, large scale serving, hot deployment and RPA configuration.

    • Staff Scientist
      • Mar 2018 - Dec 2018

      • Built a machine learning platform • Designed and implement NLP deep algorithms • Built an AI online serving system

    • United States
    • Software Development
    • 700 & Above Employee
    • Senior Software Engineer
      • May 2015 - Mar 2018

      • Built Machine Learning Infrastructure under multi-thread environment with high computational performance. • Developed a compiler of a home-made programming language for feature engineering. • Architecture design for a data platform in machine learning infrastructure, including Tensorflow serving for online scoring. • Filed patents, e.g.: "FLEXIBLE CONFIGURATION OF MODEL TRAINING PIPELINES" to build more flexible infrastructure as the major contributor. This patent is selected as… Show more • Built Machine Learning Infrastructure under multi-thread environment with high computational performance. • Developed a compiler of a home-made programming language for feature engineering. • Architecture design for a data platform in machine learning infrastructure, including Tensorflow serving for online scoring. • Filed patents, e.g.: "FLEXIBLE CONFIGURATION OF MODEL TRAINING PIPELINES" to build more flexible infrastructure as the major contributor. This patent is selected as LinkedIn Digital Signage Other patents filed with LinkedIn Colleagues: #1. "UNIFIED PARAMETER AND FEATURE ACCESS IN MACHINE LEARNING MODELS" #2. "EARLY FEEDBACK OF SCHEMATIC CORRECTNESS IN FEATURE MANAGEMENT FRAMEWORKS" Show less • Built Machine Learning Infrastructure under multi-thread environment with high computational performance. • Developed a compiler of a home-made programming language for feature engineering. • Architecture design for a data platform in machine learning infrastructure, including Tensorflow serving for online scoring. • Filed patents, e.g.: "FLEXIBLE CONFIGURATION OF MODEL TRAINING PIPELINES" to build more flexible infrastructure as the major contributor. This patent is selected as… Show more • Built Machine Learning Infrastructure under multi-thread environment with high computational performance. • Developed a compiler of a home-made programming language for feature engineering. • Architecture design for a data platform in machine learning infrastructure, including Tensorflow serving for online scoring. • Filed patents, e.g.: "FLEXIBLE CONFIGURATION OF MODEL TRAINING PIPELINES" to build more flexible infrastructure as the major contributor. This patent is selected as LinkedIn Digital Signage Other patents filed with LinkedIn Colleagues: #1. "UNIFIED PARAMETER AND FEATURE ACCESS IN MACHINE LEARNING MODELS" #2. "EARLY FEEDBACK OF SCHEMATIC CORRECTNESS IN FEATURE MANAGEMENT FRAMEWORKS" Show less

    • United States
    • Retail
    • 700 & Above Employee
    • Senior Software Engineer
      • Nov 2013 - May 2015

      • Personalization of the walmart.com website based on the user study. • Java Backend Service with Spring Config. • Model building based on the user data with Hadoop/Hive. • Dynamic page generation. • Daily based operations for the advertisement update; website personalization for the location (state/city). • Personalization of the walmart.com website based on the user study. • Java Backend Service with Spring Config. • Model building based on the user data with Hadoop/Hive. • Dynamic page generation. • Daily based operations for the advertisement update; website personalization for the location (state/city).

    • United States
    • Software Development
    • 700 & Above Employee
    • Software Engineer
      • Oct 2012 - Oct 2013

      • Implemented text feature selection with Hadoop • Using the SVM and Decision Tree Classifier for the multi-class classificaiton tasks. • Implementing sampling strategies considering the accuracy confidence. • Multi-thread programming for updating data in DynamoDB • Using Machine Learning techniques for automatic classification • Using cordigo interface for web service/client development • Implemented text feature selection with Hadoop • Using the SVM and Decision Tree Classifier for the multi-class classificaiton tasks. • Implementing sampling strategies considering the accuracy confidence. • Multi-thread programming for updating data in DynamoDB • Using Machine Learning techniques for automatic classification • Using cordigo interface for web service/client development

    • United States
    • Government Administration
    • 700 & Above Employee
    • Staff Statistician Scientist
      • Aug 2009 - Sep 2012

      • Developed widely-used high-resolution digital heart phantoms. • Developed a tool to automatically segment CTA datasets and implemented 3D image operations. Both the heart phantoms and the GUI software can be freely downloaded at https://code.google.com/archive/p/hades/ • Designed and conducted stenoses detectability study for X-ray imaging. • Performed CT dose study respect to the signal detectability. • Implemented calibration algorithms in C++ for CT reconstruction… Show more • Developed widely-used high-resolution digital heart phantoms. • Developed a tool to automatically segment CTA datasets and implemented 3D image operations. Both the heart phantoms and the GUI software can be freely downloaded at https://code.google.com/archive/p/hades/ • Designed and conducted stenoses detectability study for X-ray imaging. • Performed CT dose study respect to the signal detectability. • Implemented calibration algorithms in C++ for CT reconstruction study. • Conducted FDA case reviews for medical devices. • Developed a software to calculate SNRs based on different model observers. The source code and the user manual can be developed at https://code.google.com/archive/p/mumoc/ • CUDA coding to accelerate CT reconstruction. Show less • Developed widely-used high-resolution digital heart phantoms. • Developed a tool to automatically segment CTA datasets and implemented 3D image operations. Both the heart phantoms and the GUI software can be freely downloaded at https://code.google.com/archive/p/hades/ • Designed and conducted stenoses detectability study for X-ray imaging. • Performed CT dose study respect to the signal detectability. • Implemented calibration algorithms in C++ for CT reconstruction… Show more • Developed widely-used high-resolution digital heart phantoms. • Developed a tool to automatically segment CTA datasets and implemented 3D image operations. Both the heart phantoms and the GUI software can be freely downloaded at https://code.google.com/archive/p/hades/ • Designed and conducted stenoses detectability study for X-ray imaging. • Performed CT dose study respect to the signal detectability. • Implemented calibration algorithms in C++ for CT reconstruction study. • Conducted FDA case reviews for medical devices. • Developed a software to calculate SNRs based on different model observers. The source code and the user manual can be developed at https://code.google.com/archive/p/mumoc/ • CUDA coding to accelerate CT reconstruction. Show less

    • United States
    • Higher Education
    • 700 & Above Employee
    • Research Assistant
      • Jul 2004 - Jul 2009

      • Medical Imaging Processing • 3D Image Reconstruction with MLEM for SPECT • Optical Camera Calibration with Uniform Retro-reflective Marker Motion Tracking and Quantification • Medical Imaging Processing • 3D Image Reconstruction with MLEM for SPECT • Optical Camera Calibration with Uniform Retro-reflective Marker Motion Tracking and Quantification

    • United States
    • IT Services and IT Consulting
    • 700 & Above Employee
    • Internship
      • May 2008 - Aug 2008

      • Syntax parser to analysis CIFS and NFS protocols. • A XWindows based tool to show analysis results. • Syntax parser to analysis CIFS and NFS protocols. • A XWindows based tool to show analysis results.

    • Taiwan
    • Computer Hardware Manufacturing
    • 400 - 500 Employee
    • Software Engineer
      • Apr 2003 - Jun 2004

      • Wrote the general IDE driver for all brands of motherboards in Linux. • Coded using AT&T Assembly Language and MMX instructions for video card drivers in Linux. • Wrote the general IDE driver for all brands of motherboards in Linux. • Coded using AT&T Assembly Language and MMX instructions for video card drivers in Linux.

    • Internship
      • Jun 2002 - Dec 2002

      • Tree Animatino based on D3D. This work made me quite familiar with the graphics pipeline and CG programming for shading. • Tree Animatino based on D3D. This work made me quite familiar with the graphics pipeline and CG programming for shading.

    • Software Development
    • 100 - 200 Employee
    • Internship
      • Jun 2001 - Sep 2001

      • JPEG2000 file decode and encode. • JPEG2000 file decode and encode.

    • Nigeria
    • Food and Beverage Services
    • 200 - 300 Employee
    • Truck Loader
      • Jun 1998 - Sep 1998

      Only for money. Only for money.

Education

  • Worcester Polytechnic Institute
    Ph.D, Computer Science
    2004 - 2009
  • Zhejiang University
    Master of Science, Computer Science
    2000 - 2003
  • Zhejiang University
    Bachelor of Science, Computer Science
    1996 - 2000

Community

You need to have a working account to view this content. Click here to join now