Matthew Stern
Vice President, Software Engineering at PDT Partners- Claim this Profile
Click to upgrade to our gold package
for the full feature experience.
Topline Score
Bio
Experience
-
PDT Partners
-
United States
-
Investment Management
-
100 - 200 Employee
-
Vice President, Software Engineering
-
Jul 2020 - Present
New York City Metropolitan Area
-
-
-
Google
-
United States
-
Software Development
-
700 & Above Employee
-
Senior Software Engineer (Infrastructure)
-
Nov 2018 - Jul 2020
New York City Metropolitan Area I work on Scaffolding, an internal framework that simplifies C++ RPC server development at Google. Our largest clients are YouTube, Assistant, and Search, and we have thousands of users across the company. Our clients have strict performance requirements, so all of our features must take performance into account. Some highlights from this project: * Added a dependency injection framework, which increases developers' velocity and simplifies their code. * Added support for… Show more I work on Scaffolding, an internal framework that simplifies C++ RPC server development at Google. Our largest clients are YouTube, Assistant, and Search, and we have thousands of users across the company. Our clients have strict performance requirements, so all of our features must take performance into account. Some highlights from this project: * Added a dependency injection framework, which increases developers' velocity and simplifies their code. * Added support for gRPC-backed Scaffolding servers for both sync and async threading models. * Guided users how to use our framework and build robust Google servers
-
-
Senior Software Engineer (Ads)
-
May 2017 - Nov 2018
New York City Metropolitan Area I worked on Google Ad Manager's real-time bidding (RTB) C++ server, which retrieves third-party bids for millions of programmatic ad auctions per second. Specifically, for a particular ad auction, the server must find all interested parties, send them each a HTTP request, receive bid responses, filter invalid bids, and return the rest to the auction within hundreds of milliseconds. Highlights of this project: * Built a system to deliver auction win notifications to third parties in… Show more I worked on Google Ad Manager's real-time bidding (RTB) C++ server, which retrieves third-party bids for millions of programmatic ad auctions per second. Specifically, for a particular ad auction, the server must find all interested parties, send them each a HTTP request, receive bid responses, filter invalid bids, and return the rest to the auction within hundreds of milliseconds. Highlights of this project: * Built a system to deliver auction win notifications to third parties in real time. The system ensures 99.9% of notifications are successfully delivered and uses queuing and retrying when appropriate. * Redesigned our pipeline to ingest new ads for malware detection in real time instead of in batch processing. * Managed engineering work for critical features (e.g. compliance with the EU's GDPR laws) across five different teams * Led several projects to improve C++ code health
-
-
Senior Software Engineer (Infrastructure)
-
Jan 2016 - Jun 2017
New York City Metropolitan Area Coroner is a 20% project that ingests, analyses, and presents data from C++, Java, and Go server crashes across Google's stack. On Coroner, I built a centralized analysis service that runs custom functions in real time on each ingested coredump (C++/Go) and heapdump (Java). Google server owners write these functions and submit them to the centralized service. An example of such a function is a "heapdump reporter", which provides aggregate statistics about the Java heap at the time of GC… Show more Coroner is a 20% project that ingests, analyses, and presents data from C++, Java, and Go server crashes across Google's stack. On Coroner, I built a centralized analysis service that runs custom functions in real time on each ingested coredump (C++/Go) and heapdump (Java). Google server owners write these functions and submit them to the centralized service. An example of such a function is a "heapdump reporter", which provides aggregate statistics about the Java heap at the time of GC thrashing or OOM. The service had to pay special attention to load balancing and scalability. For example, the heapdump reporter function needs to load a multi-gigabyte file into RAM every time a Java server at Google OOMs. (I also spent time optimizing this function itself, which must efficiently aggregate across millions of nodes in the server's live object graph).
-
-
Senior Site Reliability Engineer (Docs)
-
Oct 2016 - May 2017
New York City Metropolitan Area Ensured the scalability, reliability and availability of the Google Docs suite (Docs, Sheets, Slides, Drawings and supporting infrastructure), which are all written in Java. * I implemented automated capacity tooling for our services. The tools took into account future demand, redundancy, and the geographic locations of our users and backends, acquired capacity, and updated our production footprints. * For our batch pipelines, I created reliability guidelines and standards and helped… Show more Ensured the scalability, reliability and availability of the Google Docs suite (Docs, Sheets, Slides, Drawings and supporting infrastructure), which are all written in Java. * I implemented automated capacity tooling for our services. The tools took into account future demand, redundancy, and the geographic locations of our users and backends, acquired capacity, and updated our production footprints. * For our batch pipelines, I created reliability guidelines and standards and helped debug scalability issues. In particular, I created a minified database dump that downstream pipelines used to save terabytes of wasted I/O every day. * I implemented a fault injection service for Java servers at Google, which allows service owners to simulate added latency and high error rates when conducting disaster recovery tests. * I specialized in issues related to Java GC thrashing. I tuned the GC settings and improved Java code to avoid using excessive memory. * I served on the SRE on call rotation and debugged critical issues in real time.
-
-
Site Reliability Engineer (Docs)
-
Aug 2014 - Oct 2016
New York City Metropolitan Area
-
-
-
-
Lead Software Engineer
-
Dec 2012 - Dec 2013
Ithaca, New York Area • Created, assigned and managed programming tasks for other engineers • Implemented backend to create/edit properties and user accounts in Django & MySQL • Designed and implemented a comprehensive live chat and messaging platform • Created geographical property search using Javascript and Google APIs • Integrated site with HelloSign API to allow for branded online contract signing • Implemented wireframes for frontend using HTML and CSS • Created a CDN for the website using… Show more • Created, assigned and managed programming tasks for other engineers • Implemented backend to create/edit properties and user accounts in Django & MySQL • Designed and implemented a comprehensive live chat and messaging platform • Created geographical property search using Javascript and Google APIs • Integrated site with HelloSign API to allow for branded online contract signing • Implemented wireframes for frontend using HTML and CSS • Created a CDN for the website using Amazon Cloudfront Show less
-
-
-
Amazon
-
United States
-
Software Development
-
700 & Above Employee
-
Software Developer Engineer, Intern
-
May 2013 - Aug 2013
Greater Seattle Area * Architected and implemented an autonomous, distributed and multithreaded web crawler that finds customer facing bugs on Amazon webpages and reports the issues to teams * Drafted a design document for web crawler with a proposal, full architectural overview, design tradeoffs, and performance expectations * Implemented a feed file diff algorithm that reduced by 99% the average number of database inserts that a feed file processor system makes
-
-
-
Amazon
-
United States
-
Software Development
-
700 & Above Employee
-
Software Development Engineer, Intern
-
May 2012 - Aug 2012
Greater Seattle Area * Created a scalable metrics dashboard with dynamic front end graph tools and performant file I/O persistence with Python and Django * Developed a tools dashboard website to interact with Perl web services using Perl and Mason * Developed web apps for Facebook and Twitter
-
-
Education
-
Cornell University
Bachelor of Science (B.S.), Computer Engineering -
Cornell University
Bachelor of Science (B.S.), Electrical and Computer Engineering