MATTHEW SUN Machine Learning / Data Engineer

** ******* *****, *********, ** 07045

Tel: 973-***-**** Email: adjfsn@r.postjobfree.com Linkedin: https://www.linkedin.com/in/matthew-sun-8337bb4/ Professional with 13+ years of software development, large data analytics, and end-to-end experiences to build fault-tolerant, highly scalable data streaming framework to provide real-time data-centric business solutions. Comprehensive knowledge of machine learning algorithms. Strong mathematical background with critical thinking. Extensive vendor / project management experience. SKILLS & EXPERTISE

Programming Languages: Scala Python Java SQL C++ Machine Learning: Linear Regression Ridge Lasso Logistic Regression Random Forest XGBoost PCA SVM LDA QDA Collaborative Filtering A/B Test NLP Neural Networks Distributed Computing: Kafka Spark Akka Azure Hadoop Docker UNIX Linux WORKING EXPERIENCES

Sr. Software Engineer – Machine Learning 05/2018 ~ 10/2020 Avlino - NJ

As a core team member, worked on BerthSight Management System (BMS) and YardSight Management System (YMS), which provide real-time AI business solutions to optimize logistic operations for port industry. Both BMS and YMS consist of various microservices written in scala. These microservices, built on Kafka, integrated with batch-learning capabilities from Spark, are deployed in production as docker images configured with continuous delivery pipeline in Jenkins. My major responsibilities include

Database replication: built transactional MSSQL database replication from on-premise SQL server to Azure.

Azure deployment: deployed YMS system on Azure for each client using Azure CLI script

Debezium CDC connector: wrote debezium based change data capture (CDC) connector for MySQL and MSSQL database.

Phase I: designed an Akka-based concurrent microservice, providing a query based set of current states update as input for YMS. The query was written as stored procedures. Slick, a Functional Relational Mapping (FRM) library was used to query and process data, stored procedures in MSSQL database.

Phase II: developed a streaming based subset of current states microservices, such as work-instruction, work-queue, container- sequence-input-by-slot, etc., through Kafka and Spark stream DSL join and aggregate operations. These microservices generated real-time topics as inputs for dwell-time prediction, container-sequence-optimization models, unplanned-container-decking-engine etc., as part of functional modules for BMS and YMS. Sr. Engineer – Network Engineering 11/2012 ~ 12/2017 Verizon Wireless - NJ

As OEM (Original Equipment Manufacturer) lead for Motorola and LG, helped OEMs identify top device quality issues; reduce smart phone device returns to Verizon warehouse and bring down No Trouble Found (NTF) rate on returned devices through large data analysis in Python, designing and deploying machine learning projects in Scala under Spark. The major responsibilities include

Deployed large-scale data manipulation, data mining and advanced analytical techniques to determine general and ad hoc root causes for device issues, for customer guarantee returns, etc. Provided analysis, recommendations and requirements to new product development planning teams, to OEMs and testing teams to design and test out future failures before device launches.

Recommendation System - Pioneered and participated the design and implementation of DTS (Device Trouble Shooting) guidance recommendation system for CARE.

Machine Learning - Experimented Random Forest machine learning model to predict future device return rate for Supply Chain.

Business Intelligence - Designed and integrated device performance analytics and dashboard visualization into corporate business intelligence tools.

Reduced overall device returns by 40%; overall NTF rate dropped from over 20% to single digit. Consultant – Supply Chain 05/2011 ~ 10/2012

Verizon Wireless - NJ

As a core team member working on Verizon Remote diagnostics (VRD) tool, made significant contributions towards a successful launch of VRD in 4 area call centers at VZW; played a leading role for feature management, content management, system management, diagnostics & logging, security management and report management of VRD tool. Major responsibilities include

Defined and wrote requirements for VRD and Diagnostic Agent (DA), which provided logging, analytics and diagnostics functions for tech support reps during remote. DA gradually became integrated into VRD as VRD evolved into a troubleshooting platform for CARE.

MATTHEW SUN adjfsn@r.postjobfree.com 973-***-****

2

Lead a supporting team from vendor and CARE to support VRD deployment in VZW area call centers, dramatically increasing the utilization of VRD tool at CARE.

System Engineer/Lead Architect 07/2010 ~ 04/2011

Cybercom Services Inc. – NJ (contractor position at Verizon Wireless) As a consultant, work at the Warren facility of Verizon Wireless. Major responsibilities include

Monitored and identified Android applications that presented an abnormal traffic pattern from a large pool of data aggregated from Wireless Network Guardian (WNG) tool. Designed and implemented the algorithm of how to effectively and efficiently extract and analyze the data in order of mega bytes in Python. Presented the network impact profiles of Android applications with high traffic volume to senior management.

Identified and analyzed TCP/IP protocols and algorithms used by Android applications from captured network packets. Suggested the improvement of the algorithms to keep the network traffic to its minimum to the Android application vendors. Participated in the conference call with Android application vendors to discuss the issue of traffic load and algorithm improvement. The Android applications that have been analyzed include Newsroom, Craigsnotifica, Astro File Manager, Tango Video Call, etc. Senior Software Engineer 12/2005 ~ 05/2010

Motorola Inc. - NJ

Starting from early 2009, defined and implemented feature requirements in Java on Android platform for phone application. Before migrating to Android, defined and implemented feature requirements in C++ for TAPI module on BREW platform.

Assumed the role of feature lead. Defined and implemented feature requirements, algorithms and interactions among various feature applications.

The major applications that interacted with calling include messaging, settings, frameworks, engine, multimedia, Bluetooth, phonebook, browser, camera, alarm, tools (calendar), MOD (music on demand), VOD (video on demand) and PTT (push to talk). Senior Embedded Software Engineer 07/2004 ~ 11/2005 Xtellus Inc. - NJ

Participated in the design of Dynamic Equalizing Blocker (DEB) control board module. It can selectively and dynamically pass, equalize or block any or all wavelengths on an optical fiber by software control. It is a centerpiece of Reconfigurable Optical Add/Drop Multiplexer (ROADM) in reconfigurable optical networks. It was implemented using Analog Device 21061 DSP processor in multi-threaded C and assembly with VisualDSP development tools.

Responsible for software architecture design of the DEB control board. Defined the interfaces of the different functional modules. Modules include configuration module, control module and monitor module.

Designed and implemented complicated multi-task DSP algorithms to measure, control and maintain wavelength based optical channels in real time. Each individual optical channel could be independently configured and compensated by temperature dependent equalizers.

CAREER DEVELOPMENT

Data Scientist 01/2018 ~ 04/2018

New York City Data Science Academy

An intensive 12-week data science bootcamp, providing comprehensive theory of machine learning, deep learning and large data, with hands-on projects immersed with Python, SQL, data manipulation and data visualization. EDUCATION

Master of Science in Electrical Engineering

Queen s University at Kingston, Ontario, Canada

Bachelor of Electrical Engineering

TongJi University, Shanghai, PRC

Contact this candidate