· 141 Pearl St., 4th Floor · Boston, MA, 02110 · Cell: 617-306-4212 · jforan@gmail.com ·

Jeffrey Foran

Objective

 
To be in a position where I design and oversee the construction and deployment of revolutionary computing systems.
Education

Massachusetts Institute of Technology February 1999 Cambridge, MA
 
· Masters of Engineering in Computer Science and Electrical Engineering.
· Bachelor of Science in Computer Science.
Experience

  YouTube (Google/Alphabet)
  December 2012 - Present Software Engineer (Technical Lead) San Bruno, CA
 
I have been a tech lead on the Music team for my entire tenure at YouTube. Initially, I was in charge of creating and publishing an art track for every song ever made; currently there are about 200M of them. YouTube receives DDEX album deliveries from music labels around the world which contain high quality sound recordings, cover art, and album metadata. I was responsible for creating and leveraging software systems to create these "videos" from those deliveries, and for organizing them into album playlists. The team I have been on during my tenure—of which I have been a IC TL the entire time—has been in some form or another in charge of curating and publishing tracks, releases, albums, artist metadata, and personalized music playlists. Art tracks back the functionality of the YouTube Music product.

From 2014 to 2015, I was promoted for leading two simultaneous backend migrations. The first migrated our infrastructure from batch to the newly available, globally replicated relational database (Spanner). The second migrated the layout of how YouTube's rights management system coordinated the payments for art track videos; it switched payments from a "first to deliver gets paid everything" system to a "whatever gets played gets paid" system. Without these migrations, YouTube Music would likely not have been ready for its next transition.

In 2016, Google Play Music and YouTube Music merged, and YouTube infra and branding were chosen as the "winner" and Google Play was to be deprecated. Our team changed the backend which ingested and published non-DDEX metadata from the Google Knowledge Graph to a new system we developed. The new system takes various sources of metadata and matches source entities (e.g. albums, artists) and blends the metadata, selecting the appropriate metadata for each published entity. I designed and coordinated the construction of the publishing side of the system, which provides a simple, generic mechanism for publishing the "selected" music entities to serving systems and other backends. Some of the publishing features include:
  • Automated throttling
  • Automated diffing (to determine if an outbound RPC is needed)
  • Automated dependency handling
    • e.g publishing to system A some number of seconds before publishing to system B
  • Retaining the outbound RPC metadata for debugging
In 2019, I integrated music systems with YouTube's Trust and Safety (T&S) systems, allowing for inappropriate music content to be taken down. I worked with the T&S team to ensure appropriate protections for well-known music artists and ensured that T&S policies were appropriately nuanced for music.

In 2021, we decided to move our serving system to Topic Server, the graph database that backs the Google Knowledge Graph. I have been responsible for the design of the migration, the graph database schema, the batch and real-time export of our data, and the onboarding of developers of clients from the current system to Topic Server.

In 2023, I designed how DeepMind's new music generation models would enable users to create GenAI music on the YouTube Shorts platform. In 2024, I designed the generic system infrastructure for user-initiated music and non-music GenAI media creation at YouTube.

I have also lead other smaller projects such as:
  • Publishing our music corpus metadata to Amazon Alexa
  • Designing how the music corpus would be used in YouTube Main
  • Coordinating the migration off of a deprecated art track video creation system
  Netezza/IBM
  September 2007 - Present Senior Engineer Marlborough, MA
 
Designed parallel algorithm interfaces for developers and implemented example algorithms to run on the Netezza Performance Server (NPS). Led a team that designed and developed a prototype GUI for developers and modelers working on writing and using parallel algorithms on NPS. Designed and built Python, C++, R, and Fortran language interfaces for parallel programming on NPS. Designed and oversaw development of the geospatial and bioinformatics interfaces for NPS. Implemented various trade show exhibition applications that demonstrated the capabilities of the these algorithms and interfaces. Contact: Justin Lindsey (LinkedIn)

  ITA Software
  August 2006 - September 2007 Principal Engineer Cambridge, MA
 
Designed and developed a software deployment system for the operations department of ITA. The application, Service Manager, runs on their various server machines and allows operations managers to perform complex upgrades and dynamic deployments of their various server applications from a central interface.

  LavaStorm (Purchased by Infogix; product is now Data3Sixty)
  September 2002 - August 2006 Principal Engineer Boston, MA
 
Co-designed and developed a data processing system for complex processing of large datasets. Designed solutions to generic telecom routing and billing problems. Oversaw the adaption of these generic solutions for specific customers (Verizon, Bell South, MCI). Contact: Rich Boccuzzi (LinkedIn).

  Hewlett Packard
  May 2001 - September 2002 System Architect Burlington, MA
 
Designed a distributed, web-based preview-pick-and-print system for HP marketing material. Designed and deployed a video capturing, editing, deployment, and viewing system for dynamic video viewing on handheld PCs for 650 VIPs at the Olympics. Implemented the video editing software for the same system. Designed a procurment-to-delivery process and digital asset management system for commercial print (for Virgin Megastores). Designed, co-implemented, and deployed an audio log system (for internal use at the lab). Designed, implemented, and deployed a hyperlinked video platform and built and deployed a player for the platform. Designed a distributed platform for application versioning, deployment, messaging and management. Contact: Justin Lindsey (LinkedIn) and J. T. Hwang (LinkedIn).

  LavaStorm (Purchased by Infogix)
  March 2000 - April 2001 Principal Engineer Waltham, MA
 
In charge of a team of twelve responsible for the design, implementation, and deployment of Clay Christiensen's Solution Lab - an online group-based innovation management, discussion, and publishing system. Contact: Ted Czotter (LinkedIn).

  August 1999 - March 2000 Engineer Waltham, MA
 
In charge of a sub-team of three, in a team of ten, responsible for the app-server and database design, implementation, and deployment for a windows desktop stock-ticker/news-delivery application (Infogate, later NewsTracker for USA Today). Designed and deployed the connection between Monster.com's Siebel call center and their web site. Contact: Ted Czotter (LinkedIn).

  Media Genius
  February 1999 - August 1999 Engineer Maynard, MA
 
Designed, implemented and deployed a scalable system that delivers receipts to email senders when their outgoing email is read over the web by a recipient (previously www.confirmail.com). Patent 6618747 designer and co-owner. Contact: Frank Flynn. (LinkedIn).

  Lincoln Laboratory
  February 1998 - February 1999 Research Assistant Lexington, MA
 
Designed, implemented and integrated a pronoun indentification module into a large interlingual machine translation system. The system is capable of translating many languages to and from a machine-understandable language. The module I built used various semantic and lexocographic rules during and after translation to determine the object or action to which various pronouns refer in a given section of text. This work constitutes my Masters thesis. Advisors: Cliff Weinstein and Young-Suk Lee.

  Draper Labs
  January 1997 - January 1998 Intern Cambridge, MA
 
Designed, implemented, and integrated a leak detection and gyroscope management system in a 6 foot autonomous robotic fish. Advisor: Jamie Anderson.

  MIT Nuclear Reactor
  May 1995 - January 1997 Nuclear Reactor Operator Cambridge, MA
 
Learned about the physics, fluid dynamics, infrastructure, operating procedures for the nuclear reactor at MIT. Operated the reactor and ran experiments for professors and hospitals. Trained new employees on the operation of the reactor. Contact: Edward Lau (LinkedIn).
Skills

 
· C/C++, Java, Python, R, PL/SQL, assembly. Familiar with: Perl, VB Script/ASP, Javascript, and VB. Comfortable learning new programming languages quickly.
· tls/ssl, ssh, http, smtp, pop, xml, tcp, udp, ftp, ldap, dhcp, rpc, dns, sftp, soap, uddi, mime, and others.
· Oracle, Sql Server, Netezza, and MySql.
· Design: I excel at large scale system design, and am passionate about creating systems that affect people's lives. I enjoy designing and implementing everything from low level algorithms to functional specifications to application process designs.