![[Uber Open Summit 2018] Petastorm: Training & Evaluation of Deep Learning Models @UberEngineering](https://i.ytimg.com/vi/CcGVCjqXgUY/maxresdefault.jpg)
Uploaded November 2018 | Updated September 2026, 2 weeks ago
Petastorm is an open source data access library that enables single machine or distributed training and evaluation of deep learning models directly from datasets in Apache Parquet format. Petastorm supports popular Python-based machine learning frameworks such as Tensorflow, Pytorch, and PySpark. In this talk during Uber Open Summit 2018, Pyro lead Yevgeni Litvin presents Petastorm features and show how Petastorm shortens the model development cycle at Uber ATG.
Learn more about Uber Open Source: uber.github.io
Petastorm is an open source data access library that enables single machine or distributed training and evaluation of deep learning models directly from datasets in Apache Parquet format. Petastorm supports popular Python-based machine learning frameworks such as Tensorflow, Pytorch, and PySpark. In this talk during Uber Open Summit 2018, Pyro lead Yevgeni Litvin presents Petastorm features and show how Petastorm shortens the model development cycle at Uber ATG.
Learn more about Uber Open Source: uber.github.io
![[Uber Open Summit 2018] Petastorm: Training & Evaluation of Deep Learning Models](https://i.ytimg.com/vi/CcGVCjqXgUY/mqdefault.jpg "[Uber Open Summit 2018] Petastorm: Training & Evaluation of Deep Learning Models
Petastorm is an open source data access library that enables single machine or distributed training and evaluation of deep learning models directly from datasets in Apache Parquet format. Petastorm supports popular Python-based machine learning frameworks such as Tensorflow, Pytorch, and PySpark. In this talk during Uber Open Summit 2018, Pyro lead Yevgeni Litvin presents Petastorm features and show how Petastorm shortens the model development cycle at Uber ATG.
Learn more about Uber Open Source: https://uber.github.io/")
![[Uber Open Source] Apache Zeppelin: Preview 0.9.0](https://i.ytimg.com/vi/CgLTZuyB2a0/mqdefault.jpg "[Uber Open Source] Apache Zeppelin: Preview 0.9.0
During this April 2019 meetup in San Francisco, Moon Soo Lee, co-founder of ZEPL, introduces Apache Zeppelin 0.9.0 and discusses key features for integrating machine learning models into data science notebooks using Zeppelin.")
![[StackUp] vis.gl: Open source Tools for Large-scale Geospatial Data Visualization](https://i.ytimg.com/vi/Ci7nwuAfX_o/mqdefault.jpg "[StackUp] vis.gl: Open source Tools for Large-scale Geospatial Data Visualization
During this July 2019 meetup at Uber’s New York City office, Uber engineer Tarek Sherif presents vis.gl, Ubers open source suite of geospatial data visualization frameworks.")
![[Uber Open Source ] Distributed Deep Learning with Horovod Alex Sergeev](https://i.ytimg.com/vi/D1By2hy4Ecw/mqdefault.jpg "[Uber Open Source ] Distributed Deep Learning with Horovod Alex Sergeev
During a March 2019 meetup hosted at Ubers SF office, Uber senior software engineer Alex Sergeev discusses how frameworks like TensorFlow, PyTorch, and MXNet make it easy to design and train deep learning models. However, when it comes to scaling models to multiple GPUs in a server, or multiple servers in a cluster, difficulties usually arise.")
![[Uber Open Summit 2018] Scaling a React Component Library: Building Uber’s new Design Language](https://i.ytimg.com/vi/DmHlOvEGdfk/mqdefault.jpg "[Uber Open Summit 2018] Scaling a React Component Library: Building Uber’s new Design Language
Duplication of code and inconsistent design and experience are major issues across web applications. During this Uber Open Summit 2018 tech talk, engineer Nadiia Dmytrenko discusses the process of designing and building a high-quality library of reusable, composable, themeable, and highly-customizable React component library to create a consistent design.")
![[Destination:Web] Discovering Web Audio by Recreating Deep Note](https://i.ytimg.com/vi/DtqTEZJHeXU/mqdefault.jpg "[Destination:Web] Discovering Web Audio by Recreating Deep Note
During a January 2019 meetup, Facebook software engineer Stoyan Stefanov discusses how to recreate the THXs Deep Note, an audio logo played before most movies start, in JavaScript using the Web Audio API.")
![[Payments Platform] Evolution of Payments at Uber Nimish Sheth & Steven Karis](https://i.ytimg.com/vi/Dz6dAZs8Scg/mqdefault.jpg "[Payments Platform] Evolution of Payments at Uber Nimish Sheth & Steven Karis
Uber’s payments architecture is composed of two main parts: collections and disbursements. In this presentation from a September 2018 meetup, engineers Nimish Sheth & Steven Karis offers a closer look at our high-level payments stack, core data models, and cash money movements.")
![[Uber Open Source] Leveraging Pinot at Uber for Large-scale Analytics](https://i.ytimg.com/vi/E8Xsmc9AqDM/mqdefault.jpg "[Uber Open Source] Leveraging Pinot at Uber for Large-scale Analytics
Apache Pinot is a distributed columnar database for real-time analytics. It powers many well-known applications, including LinkedIn’s Who Viewed My Profile feature and Uber Eats timely analytics. Since Pinot joined Apache Incubator in 2019, it has gained a lot of momentum with a growing community.
During this May 2020 Uber Engineering Meetup, software engineers Haibo Wang and James Shao discuss how Pinot empowers timely analytics at Uber. We will also dive into how we built an open source Presto connector to enable complex SQL queries on Pinot and other data sources, as well as how we developed upserts for Pinot to allow mutable data in real-time tables.
Learn more: https://eng.uber.com/engineering-sql-support-on-apache-pinot/")
![[Uber Eng Blog] Visualizing Uber Air Simulated Flights](https://i.ytimg.com/vi/Edr0Adlka-Y/mqdefault.jpg "[Uber Eng Blog] Visualizing Uber Air Simulated Flights
In this visualization from the Tuesday, October 29, 2019 Uber Engineering Blog article, Taking City Visualization into the Third Dimension with Point Clouds, 3D Tiles, and deck.gl, Uber Air flight routes are simulated using our visualization suite.
Learn more about these powerful data visualizations tools on the Uber Engineering Blog: eng.uber.com")
![[Go NYC] The Myth around Go Frameworks & Libraries Venkatesh Thallam](https://i.ytimg.com/vi/EiBHiMexzEI/mqdefault.jpg "[Go NYC] The Myth around Go Frameworks & Libraries Venkatesh Thallam
During an October 2018 Go Language NYC meetup, Paperless Post engineer Venkatesh Thallam discusses apprehensions using Go frameworks and external libraries and offers suggestions to increase Go development flexibility.")
