|
name
|
viewport
|
width=device-width, initial-scale=1.0
|
|
name
|
color-scheme
|
light dark
|
|
name
|
author
|
John Myles White
|
|
name
|
description
|
Start
End
Role
Description
2026
Present
Starting Something New
I’m exploring application areas where AI hasn’t yet made inroads.
2013
2026
Meta
I started as an IC4 PhD new grad and left as a D1 engineering director.
2025
2026
Meta :: MSL Infra
I supported an org called PyTorch Systems, focused on developing libraries and infra for PyTorch, including torchtune, PyTorch’s internal DevInfra org, PyTorch’s Distributed Infra group and more.
2024
2025
Meta :: AI Infra
I supported an org called PyTorch Components, focused on PyTorch libraries like checkpointing, TorchRec and also many Meta-wide Content Understanding models, including the text and speech translation models, text, image and video embeddings models. Before PyTorch, I supported the F3 Compilers group, who maintained an ML engineering feature definition and computation engine for Ads.
2019
2024
Meta :: Data Infra
I moved to Data Infra to support all of the various query language efforts before taking on managing other infra systems including Chronos, Meta’s scheduling system over a large machine fleet for scheduled jobs; Dataswarm, the Meta orchestration system that inspired Apache Airflow; and UPM, a query analyzer and SQL extension system that was the inspiration for the SDF startup.
2018
2019
Meta :: DevInfra
I moved over to DevInfra to manage the Python SWE team. We worked on things like performance optimizations of async/await and Strict Python.
2013
2018
Meta :: Data Science
I joined Cameron Marlow’s Data Science team right as he left and the team was split into separate Core Data Science and Data Science Infrastructure teams. I was part of Core Data Science, but I worked on Data Science Infrastructure projects much of the time. I started managing in 2017 after moving back home to NYC. I worked on systems like Bento, Meta’s hosted notebook system; Deltoid, Meta’s A/B testing platform; and Perflab, a benchmarking tool for comparing the performance of two versions of HHVM with each other.
2013
2013
MIT CSAIL
I was a research affiliate working on the Julia programming language after finishing my PhD before going to Facebook.
2004
2008
Web Developer
I worked as a web developer at various small firms (e.g. Connors Communications) while playing semi-professionally in a few different bands.
|
|
name
|
keywords
|
blog,personal,programming,statistics,data
|
|
name
|
twitter:card
|
summary
|
|
name
|
twitter:title
|
Work
|
|
name
|
twitter:description
|
Start End Role Description 2026 Present Starting Something New I’m exploring application areas where AI hasn’t yet made inroads. 2013 2026 Meta I started as an IC4 PhD new grad and left as a D1 engineering director. 2025 2026 Meta :: MSL Infra I supported an org called PyTorch Systems, focused on developing libraries and infra for PyTorch, including torchtune, PyTorch’s internal DevInfra org, PyTorch’s Distributed Infra group and more. 2024 2025 Meta :: AI Infra I supported an org called PyTorch Components, focused on PyTorch libraries like checkpointing, TorchRec and also many Meta-wide Content Understanding models, including the text and speech translation models, text, image and video embeddings models. Before PyTorch, I supported the F3 Compilers group, who maintained an ML engineering feature definition and computation engine for Ads. 2019 2024 Meta :: Data Infra I moved to Data Infra to support all of the various query language efforts before taking on managing other infra systems including Chronos, Meta’s scheduling system over a large machine fleet for scheduled jobs; Dataswarm, the Meta orchestration system that inspired Apache Airflow; and UPM, a query analyzer and SQL extension system that was the inspiration for the SDF startup. 2018 2019 Meta :: DevInfra I moved over to DevInfra to manage the Python SWE team. We worked on things like performance optimizations of async/await and Strict Python. 2013 2018 Meta :: Data Science I joined Cameron Marlow’s Data Science team right as he left and the team was split into separate Core Data Science and Data Science Infrastructure teams. I was part of Core Data Science, but I worked on Data Science Infrastructure projects much of the time. I started managing in 2017 after moving back home to NYC. I worked on systems like Bento, Meta’s hosted notebook system; Deltoid, Meta’s A/B testing platform; and Perflab, a benchmarking tool for comparing the performance of two versions of HHVM with each other. 2013 2013 MIT CSAIL I was a research affiliate working on the Julia programming language after finishing my PhD before going to Facebook. 2004 2008 Web Developer I worked as a web developer at various small firms (e.g. Connors Communications) while playing semi-professionally in a few different bands.
|
|
property
|
og:url
|
https://www.johnmyleswhite.com/work/
|
|
property
|
og:site_name
|
John Myles White
|
|
property
|
og:title
|
Work
|
|
property
|
og:description
|
Start End Role Description 2026 Present Starting Something New I’m exploring application areas where AI hasn’t yet made inroads. 2013 2026 Meta I started as an IC4 PhD new grad and left as a D1 engineering director. 2025 2026 Meta :: MSL Infra I supported an org called PyTorch Systems, focused on developing libraries and infra for PyTorch, including torchtune, PyTorch’s internal DevInfra org, PyTorch’s Distributed Infra group and more. 2024 2025 Meta :: AI Infra I supported an org called PyTorch Components, focused on PyTorch libraries like checkpointing, TorchRec and also many Meta-wide Content Understanding models, including the text and speech translation models, text, image and video embeddings models. Before PyTorch, I supported the F3 Compilers group, who maintained an ML engineering feature definition and computation engine for Ads. 2019 2024 Meta :: Data Infra I moved to Data Infra to support all of the various query language efforts before taking on managing other infra systems including Chronos, Meta’s scheduling system over a large machine fleet for scheduled jobs; Dataswarm, the Meta orchestration system that inspired Apache Airflow; and UPM, a query analyzer and SQL extension system that was the inspiration for the SDF startup. 2018 2019 Meta :: DevInfra I moved over to DevInfra to manage the Python SWE team. We worked on things like performance optimizations of async/await and Strict Python. 2013 2018 Meta :: Data Science I joined Cameron Marlow’s Data Science team right as he left and the team was split into separate Core Data Science and Data Science Infrastructure teams. I was part of Core Data Science, but I worked on Data Science Infrastructure projects much of the time. I started managing in 2017 after moving back home to NYC. I worked on systems like Bento, Meta’s hosted notebook system; Deltoid, Meta’s A/B testing platform; and Perflab, a benchmarking tool for comparing the performance of two versions of HHVM with each other. 2013 2013 MIT CSAIL I was a research affiliate working on the Julia programming language after finishing my PhD before going to Facebook. 2004 2008 Web Developer I worked as a web developer at various small firms (e.g. Connors Communications) while playing semi-professionally in a few different bands.
|
|
property
|
og:locale
|
en
|
|
property
|
og:type
|
article
|
Apache