Dezhi “Andy” Fang
(802) 294-3264 ·
andy@andyfang.me ·
New York, NY
Full CV: https://cv.andyfang.me ·
Software Engineer
Georgia Institute of Technology
Major in Computer Science, Information Network & Intelligence. GPA: 3.54/4.0.
Software Engineer
Enterprise Data Platform. Technical lead of trading-critical security master, providing timely and reliable access to fundamental reference datasets at Citadel.
- Architected and led the implementation of observability stack to enable consistent monitoring of service performance and health.
- Integrated distributed tracing, unified metrics, dashboard, and alerting into several key systems. Set widely-adopted observability strategies.
- Designed and built organization-wide Kubernetes deployments tooling and best practices.
- Designed and led build-out of firm-wide, Spinnaker-based continuous deployment system for both Kubernetes and legacy NFS deployments.
- Modernized of build system for large-scale C++ (migration to Bazel) and Java (multi-project Gradle monorepo builds) services.
- Optimized performance and memory usage for large-scale C++ services, such as identifying memory fragmentation in glibc
malloc()
, which resulted in dramatic decrease in OOM and more than 75% memory usage reduction for a critical production service.
Software Development Engineer
May, 2017 to Aug., 2017 (intern)
Cloud Infrastructure, Production Platform. Select projects include:
- Designed and led the build-out of Airbnb's distributed, delayed job scheduling and queueing system on top of AWS SQS and DynamoDB.
- Read more about this work on Airbnb Engineering Blog: medium.com/p/93ab10f05f99.
- Led an intern as intern manager during Summer 2019 for this work.
- Designed and implemented multi-cluster architecture for Airbnb's Kafka clusters, including a metadata proxy that multiplexes Kafka requests based on the Kafka TCP protocol.
- Implemented a thrift serializer/deserializer library for ruby that's more than 25x faster than the existing binding.
- Read more about this work on Airbnb Engineering Blog: andf.us/sparsam. Find this work open-sourced at: github.com/airbnb/sparsam.
- Built distributed tracing pipeline for collecting frontend performance metrics.
- Built performance profiling and monitoring tooling for JVM applications running in Kuberenetes/Docker.
Undergraduate Research Assistant
Research in the cross-section of data mining and human-computer interaction with professor Polo Chau.
- One of the investigators in NIH's MD2K (Mobile Data to Knowledge) initiative. Developed a predictive visualization dashboard for exploring correlations between activity, stress, and smoking relapses. Paper.
- MMap: Scaling up scientific computation with memory mapping. Paper. ARGO: Billion-scale visualization of network data. Paper.
STAR Intern on Software Engineering
Utilizing data mining (with Hadoop and Python) to improve malware detection efficacy.
- Inferring unknown files' reputations with parent file and execution environment features.
- Deployed new rule in Symantec's AV Engine with a false positive rate of only 0.2%. Used in Norton Security and Symantec Endpoint Protection.
- Implemented a decision tree executor for detecting malware using file features that replaced Symantec's existing implementation.
Dynein: Building an Open-source Distributed Delayed Job Queueing System [andf.us/dynein]
Airbnb Engineering & Data Science Blog
Short URL: andf.us/dynein, https://medium.com/p/93ab10f05f99
Accelerating Services at Airbnb by Building a Blazing Fast Thrift Binding for Ruby [andf.us/sparsam]
Airbnb Engineering & Data Science Blog
mHealth Visual Discovery Dashboard [PDF][Video]
Demo, ACM International Joint Conference on Pervasive and Ubiquitous Computing (UBICOMP)
Carina: Interactive Million-Node Graph Visualization using Web Browser Technologies [arXiv][PDF]
The 2017 World Wide Web conference (WWW’17)
[Presenter] M-Flash: Fast Billion-scale Graph Computation Using a Bimodal Block Processing Model [Slides][arXiv][PDF]
Joint European Conference on Machine Learning and Knowledge Discovery in Databases (ECML-PKDD’16)
Visual Exploration of Machine Learning Results using Data Cube Analysis [ACM][PDF]
Proceedings of the Workshop on Human-In-the-Loop Data Analytics (HILDA’16) (Co-located with SIGMOD/PODS’16)
M3: Scaling Up Machine Learning via Memory Mapping [arXiv][PDF]
Proceedings of the 2016 International Conference on Management of Data (SIGMOD/PODS’16)
Outstanding Undergraduate Researcher Award
Undergraduate Research Opportunities Program, Georgia Institute of Technology
Annually, one student from each department in Georgia Tech is nominated to receive the award on the basis of involvement in long-term research projects, participation in conferences, published research papers, displayed leadership within the research environment, and unique contributions to the field.
Outstanding Undergraduate Research
College of Computing, Georgia Institute of Technology
Annually, one student from College of Computing is nominated by faculty members to receive the award, recognizing their involvement in undergraduate research.
Material, Supplies and Technology Grants (MS&T)
Georgia Institute of Technology
Supported by MS&T grants for research in Visualization and Virtual Reality during Summer 2017. $1,000.
President's Undergraduate Research Award (PURA)
Georgia Institute of Technology
The Undergraduate Research Opportunities Program competitively funds individual requests by a student to support undergraduate student involvement in faculty research. Dezhi “Andy” Fang received:
- PURA Travel Award funding to present at 2016 ACM SIGMOD/PODS @ San Francisco, USA. $500;
- PURA Salary Award (PURA) for faculty research in Fall 2016, $1,500.
Undergraduate Research Poster Competition Finalist
ACM Special Interest Group on Management of Data (SIGMOD 2016)
Presented M3: Scaling Up Machine Learning via Memory Mapping. Awarded to ~10 recipients globally each year.
Georgia Tech Online Masters Program Scholarship
Georgia Tech College Of Computing
A full scholarship of the Georgia Tech Online Masters Program (worth $7000) was awarded.
Third Prize, China Adolescents Science & Technology Innovation Contest
China Association for Science and Technology
Presented a stereo in-door location system using computer vision.
First Prize, National Olympiad in Informatics in Provinces
China Computer Federation (CCF)
Programming competition in a form similar to ACM-ICPC.
Georgia Tech to Participate for the First Time in Supercomputing Student Cluster Competition [article]
The Compiler (Georgia Tech College of Computing)
While these students are still in their undergrad, the training and education they have received in preparation for this competition expands well beyond HPC and produces an educational foundation they will continue using for years to come.
China Economy Draws More Students Back From Abroad [article]
The Wall Street Journal
‘The U.S. is still the greatest place for doing cutting-edge research.’ —Dezhi Fang, a junior at the Georgia Institute of Technology
Companies Cash In On Us Universities' Video Interview Requirements [article]
China Daily
Runner Up, Emory GCC Case Competition
Emory Global China Connection
Case competition focused on mobile payment industry. Developed strategy for a U.S. based company to enter China's market with NFC payment technology.
Runner Up, BizTech'16 Case Competition
Georgia Tech Management Information Systems (MIS) Club
Case competition focused on Electronic Health Record (EHR) industry. Developed strategy for an established EHR company to transform its infrastructure to cloud-based solutions with high interoperability among EHR standards and medical devices.
Runner Up in Education Track, HackDuke'15
HackDuke
Developed ResumeWorks, a platform for creating different versions of resumes to fit each job application.
Hacker's Choice Award, VandyHacks'15
VandyHacks
Developed Bank Guard, an app for detecting credit card frauds.
4th Place, IronCoder'15
Cardlytics
A semiannually held programming competition by Cardlytics.
First Place, Georgia Tech College of Computing Alumni Hackathon'15
Georgia Tech College Of Computing
Developed MapsOnPoint, an app for finding points of interest along your road trip route without taking an unreasonable detour.
django-asyncmailer
- An async email sending solution with load balancing and routing among multiple SMTP credentials
- Now sending 100k+ emails per month with a delivery rate of more than 98%
ResumeWorks
- 2nd Best Hack of Education in HackDuke
- A tool for creating different versions of resumes to fit each job application
- User can choose a subset of their experiences to match with the company's interest
Maps OnPoint
- First Place in Hackathon @ Georgia Tech College of Computing
- Web app for planning trip navigation with a quick stop for food at any point along the route
Programming Languages: Python, Java, Scala, Bash, C/C++, JavaScript/TypeScript, Ruby
Infrastructure: Apache Kafka, Kubernetes, Docker, Chef, Terraform, Amazon AWS, Netty, Redis, MySQL, DynamoDB, SQS, Resque, Quartz
Data Mining: Flink, Spark, Hive/Presto, Jupyter, Pandas, NumPy, Scikit-learn
Web Development: Django, NodeJS, Webpack, React, Three.js, D3.js