Anyscale's guest appearances

Anyscale

The Python Podcast.__init__

Machine Learning Archives - Software Engineering Daily

Intel on AI

OwlTail

Summary Building a machine learning application is inherently complex. Once it becomes necessary to scale the operation or training of the model, or introduce online re-training the process becomes even more challenging. In order to reduce the operational burden of AI developers Robert Nishihara helped to create the Ray framework that handles the distributed computing aspects of machine learning operations. To support the ongoing development and simplify adoption of Ray he co-founded Anyscale. In this episode he re-joins the show to share how the project, its community, and the ecosystem around it have grown and evolved over the intervening two years. He also explains how the techniques and adoption of machine learning have influenced the direction of the project. Announcements  Hello and welcome to Podcast.__init__, the podcast about Python’s role in data and science. When you’re ready to launch your next app or want to try a project you hear about on the show, you’ll need somewhere to deploy it, so take a look at our friends over at Linode. With the launch of their managed Kubernetes platform it’s easy to get started with the next generation of deployment and scaling, powered by the battle tested Linode platform, including simple pricing, node balancers, 40Gbit networking, dedicated CPU and GPU instances, and worldwide data centers. Go to pythonpodcast.com/linode and get a $100 credit to try out a Kubernetes cluster of your own. And don’t forget to thank them for their continued support of this show! Your host as usual is Tobias Macey and today I’m interviewing Robert Nishihara about his work at Anyscale and the Ray distributed execution framework  Interview  Introductions How did you get introduced to Python? Can you describe what Anyscale is and the story behind it? How has the Ray project and ecosystem evolved since we last spoke? (2 years ago) How has the landscape of AI/ML technologies and techniques shifted in that time?   What are the main areas where organizations are trying to apply ML/AI? What are some of the issues that teams encounter when trying to move from prototype to production with ML/AI applications? What are the features of Ray that help to mitigate those challenges?   With the introduction of more widely available streaming/real-time technologies the viability of reinforcement learning has increased. What new challenges does that approach introduce? What are some of the operational complexities associated with managing a deployment of Ray? What are some of the specialized utilities that you have had to develop to maintain a large and multi-tenant platform for your customers?   What is the governance model around the Ray project and how does the work at Anyscale influence the roadmap? What are the most interesting, innovative, or unexpected ways that you have seen Anyscale/Ray used? What are the most interesting, unexpected, or challenging lessons that you have learned while working on Ray and Anyscale? When is Anyscale/Ray the wrong choice? What do you have planned for the future of Anyscale/Ray?  Keep In Touch  robertnishihara on GitHub @robertnishihara on Twitter Website LinkedIn  Picks  Tobias The Edge Chronicles: Beyond The Deepwoods   Robert Production RL Summit Project Hail Mary by Andy Weir    Closing Announcements  Thank you for listening! Don’t forget to check out our other show, the Data Engineering Podcast for the latest on modern data management. Visit the site to subscribe to the show, sign up for the mailing list, and read the show notes. If you’ve learned something or tried out a project from the show then tell us about it! Email hosts@podcastinit.com) with your story. To help other people find the show please leave a review on iTunes and tell your friends and co-workers  Links  Ray  Podcast Episode   Anyscale UC Berkeley Matlab Deep Learning Pandas NumPy Horovod  Podcast Episode   XGBoost Modin  Podcast Episode   Dask Ray Datasets Reinforcement Learning Production Reinforcement Learning Summit AlphaGo Databricks Snowflake  Data Engineering Podcast Episode   TPU == Tensor Processing Unit Weights and Biases MLFlow RLLib Ray Serve  The intro and outro music is from Requiem for a Fish The Freak Fandango Orchestra  / CC BY-SA

<h2>Summary</h2> <p>Building a machine learning application is inherently complex. Once it becomes necessary to scale the operation or training of the model, or introduce online re-training the process becomes even more challenging. In order to reduce the operational burden of AI developers Robert Nishihara helped to create the Ray framework that handles the distributed computing aspects of machine learning operations. To support the ongoing development and simplify adoption of Ray he co-founded Anyscale. In this episode he re-joins the show to share how the project, its community, and the ecosystem around it have grown and evolved over the intervening two years. He also explains how the techniques and adoption of machine learning have influenced the direction of the project.</p> <h2>Announcements</h2> <ul> <li>Hello and welcome to Podcast.__init__, the podcast about Python’s role in data and science.</li> <li>When you’re ready to launch your next app or want to try a project you hear about on the show, you’ll need somewhere to deploy it, so take a look at our friends over at Linode. With the launch of their managed Kubernetes platform it’s easy to get started with the next generation of deployment and scaling, powered by the battle tested Linode platform, including simple pricing, node balancers, 40Gbit networking, dedicated CPU and GPU instances, and worldwide data centers. Go to <a href="https://www.pythonpodcast.com/linode?utm_source=rss&amp;utm_medium=rss" rel="nofollow" target="_blank">pythonpodcast.com/linode</a> and get a $100 credit to try out a Kubernetes cluster of your own. And don’t forget to thank them for their continued support of this show!</li> <li>Your host as usual is Tobias Macey and today I’m interviewing Robert Nishihara about his work at Anyscale and the Ray distributed execution framework</li> </ul> <h2>Interview</h2> <ul> <li>Introductions</li> <li>How did you get introduced to Python?</li> <li>Can you describe what Anyscale is and the story behind it?</li> <li>How has the Ray project and ecosystem evolved since we last spoke? (2 years ago)<br><ul> <li>How has the landscape of AI/ML technologies and techniques shifted in that time?</li> </ul> </li> <li>What are the main areas where organizations are trying to apply ML/AI?</li> <li>What are some of the issues that teams encounter when trying to move from prototype to production with ML/AI applications?<br><ul> <li>What are the features of Ray that help to mitigate those challenges?</li> </ul> </li> <li>With the introduction of more widely available streaming/real-time technologies the viability of reinforcement learning has increased. What new challenges does that approach introduce?</li> <li>What are some of the operational complexities associated with managing a deployment of Ray?<br><ul> <li>What are some of the specialized utilities that you have had to develop to maintain a large and multi-tenant platform for your customers?</li> </ul> </li> <li>What is the governance model around the Ray project and how does the work at Anyscale influence the roadmap?</li> <li>What are the most interesting, innovative, or unexpected ways that you have seen Anyscale/Ray used?</li> <li>What are the most interesting, unexpected, or challenging lessons that you have learned while working on Ray and Anyscale?</li> <li>When is Anyscale/Ray the wrong choice?</li> <li>What do you have planned for the future of Anyscale/Ray?</li> </ul> <h2>Keep In Touch</h2> <ul> <li><a href="https://github.com/robertnishihara?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">robertnishihara</a> on GitHub</li> <li><a href="https://twitter.com/robertnishihara?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">@robertnishihara</a> on Twitter</li> <li><a href="http://www.robertnishihara.com/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Website</a></li> <li><a href="https://www.linkedin.com/in/robert-nishihara-b6465444/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">LinkedIn</a></li> </ul> <h2>Picks</h2> <ul> <li>Tobias<br><ul> <li><a href="https://edgechronicles.co.uk/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">The Edge Chronicles</a>: <a href="http://www.edgechronicles.co.uk/discover-the-books/the-twig-saga/beyond-the-deepwoods/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Beyond The Deepwoods</a></li> </ul> </li> <li>Robert<br><ul> <li><a href="https://www.anyscale.com/production-rl-summit?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Production RL Summit</a></li> <li><a href="https://www.andyweirauthor.com/books/project-hail-mary-hc/project-hail-mary-el?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Project Hail Mary</a> by Andy Weir</li> </ul> </li> </ul> <h2>Closing Announcements</h2> <ul> <li>Thank you for listening! Don’t forget to check out our other show, the <a href="https://www.dataengineeringpodcast.com?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Data Engineering Podcast</a> for the latest on modern data management.</li> <li>Visit the <a href="https://www.pythonpodcast.com?utm_source=rss&amp;utm_medium=rss" rel="nofollow" target="_blank">site</a> to subscribe to the show, sign up for the mailing list, and read the show notes.</li> <li>If you’ve learned something or tried out a project from the show then tell us about it! Email <a href="mailto:hosts@podcastinit.com">hosts@podcastinit.com</a>) with your story.</li> <li>To help other people find the show please leave a review on <a href="https://itunes.apple.com/us/podcast/podcast.-init/id981834425?mt=2&amp;uo=6&amp;at=&amp;ct=&amp;utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">iTunes</a> and tell your friends and co-workers</li> </ul> <h2>Links</h2> <ul> <li><a href="https://docs.ray.io/en/latest/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Ray</a> <ul> <li><a href="https://www.pythonpodcast.com/ray-distributed-computing-episode-258/?utm_source=rss&amp;utm_medium=rss" rel="nofollow" target="_blank">Podcast Episode</a></li> </ul> </li> <li><a href="https://www.anyscale.com/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Anyscale</a></li> <li><a href="https://www.berkeley.edu/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">UC Berkeley</a></li> <li><a href="https://www.mathworks.com/products/matlab.html?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Matlab</a></li> <li><a href="https://en.wikipedia.org/wiki/Deep_learning?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Deep Learning</a></li> <li><a href="https://pandas.pydata.org/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Pandas</a></li> <li><a href="https://numpy.org/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">NumPy</a></li> <li><a href="https://horovod.ai/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Horovod</a> <ul> <li><a href="https://www.pythonpodcast.com/ludwig-horovod-distributed-declarative-deep-learning-episode-341/?utm_source=rss&amp;utm_medium=rss" rel="nofollow" target="_blank">Podcast Episode</a></li> </ul> </li> <li><a href="https://xgboost.ai/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">XGBoost</a></li> <li><a href="https://modin.readthedocs.io/en/latest/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Modin</a> <ul> <li><a href="https://www.pythonpodcast.com/modin-parallel-dataframe-episode-324/?utm_source=rss&amp;utm_medium=rss" rel="nofollow" target="_blank">Podcast Episode</a></li> </ul> </li> <li><a href="https://dask.org/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Dask</a></li> <li><a href="https://docs.ray.io/en/latest/data/dataset.html?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Ray Datasets</a></li> <li><a href="https://en.wikipedia.org/wiki/Reinforcement_learning?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Reinforcement Learning</a></li> <li><a href="https://www.anyscale.com/production-rl-summit?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Production Reinforcement Learning Summit</a></li> <li><a href="https://en.wikipedia.org/wiki/AlphaGo?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">AlphaGo</a></li> <li><a href="https://databricks.com/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Databricks</a></li> <li><a href="https://www.snowflake.com/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Snowflake</a> <ul> <li><a href="https://www.dataengineeringpodcast.com/snowflakedb-cloud-data-warehouse-episode-110/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Data Engineering Podcast Episode</a></li> </ul> </li> <li><a href="https://en.wikipedia.org/wiki/Tensor_Processing_Unit?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">TPU == Tensor Processing Unit</a></li> <li><a href="https://www.wandb.com/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Weights and Biases</a></li> <li><a href="https://mlflow.org/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">MLFlow</a></li> <li><a href="https://docs.ray.io/en/latest/rllib/index.html?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">RLLib</a></li> <li><a href="https://docs.ray.io/en/latest/serve/index.html?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">Ray Serve</a></li> </ul> <p>The intro and outro music is from Requiem for a Fish <a href="http://freemusicarchive.org/music/The_Freak_Fandango_Orchestra/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">The Freak Fandango Orchestra</a>  / <a href="http://creativecommons.org/licenses/by-sa/3.0/?utm_source=rss&amp;utm_medium=rss" rel="noopener nofollow" target="_blank">CC BY-SA</a></p>

Accelerate The Development And Delivery Of Your Machine Learning Applications Using Ray And Deploy It At Anyscale

In this Intel on AI podcast episode: Today, the deluge of data has made demand for machine learning engineers explode. Also because distributed computing is a challenging and elite subfield of computer programming, finding engineers to address these skill sets can be even more challenging and limit many business from being able to take advantage of advanced technologies like machine learning (ML). Dean Wampler, the Head of Developer Relations at Anyscale, joins the Intel on AI podcast to talk about how the Ray framework, which is heavily developed and supported by Anyscale, enables any developer to easily write distributed applications which are performant, debuggable, and maintainable. He illustrates how Ray helps developers, enterprises and organizations solve their problems without having to worry about scalable infrastructure and without needing to be experts in distributed computing. Dean discusses some of the biggest users of Ray utilize it to support their infrastructure especially during incredibly high traffic volume events to do general processes, payment processing, and fraud detection. He also describes how other companies are using Ray to do reinforcement learning and business process automation. Lastly, Dean talks about how many teams within Intel are leveraging the Ray framework for model training and reinforcement learning and at the same time working together with Anyscale to contribute to Ray and optimize it for Intel architecture. Lastly, Dean mentioned that in light of growing concerns about COVID-19, they have decided to postpone Ray Summit to late Summer or early Fall of 2020. To learn more, visit: anyscale.io Visit Intel AI Builders at: builders.intel.com/ai

<p>In this Intel on AI podcast episode: Today, the deluge of data has made demand for machine learning engineers explode. Also because distributed computing is a challenging and elite subfield of computer programming, finding engineers to address these skill sets can be even more challenging and limit many business from being able to take advantage of advanced technologies like machine learning (ML). Dean Wampler, the Head of Developer Relations at Anyscale, joins the Intel on AI podcast to talk about how the Ray framework, which is heavily developed and supported by Anyscale, enables any developer to easily write distributed applications which are performant, debuggable, and maintainable. He illustrates how Ray helps developers, enterprises and organizations solve their problems without having to worry about scalable infrastructure and without needing to be experts in distributed computing. Dean discusses some of the biggest users of Ray utilize it to support their infrastructure especially during incredibly high traffic volume events to do general processes, payment processing, and fraud detection. He also describes how other companies are using Ray to do reinforcement learning and business process automation. Lastly, Dean talks about how many teams within Intel are leveraging the Ray framework for model training and reinforcement learning and at the same time working together with Anyscale to contribute to Ray and optimize it for Intel architecture. Lastly, Dean mentioned that in light of growing concerns about COVID-19, they have decided to postpone Ray Summit to late Summer or early Fall of 2020.</p> <p>To learn more, visit: <a href="http://anyscale.io" rel="nofollow" target="_blank">anyscale.io</a></p> <p>Visit Intel AI Builders at: <a href="http://builders.intel.com/ai" rel="nofollow" target="_blank">builders.intel.com/ai</a></p>

Making Machine Learning Application Development Easy with Ray and Anyscale – Intel on AI – Episode 50

Ion Stoica

Machine learning applications are widely deployed across the software industry.  Most of these applications used supervised learning, a process in which labeled data sets are used to find correlations between the labels and the trends in that underlying data. But supervised learning is only one application of machine learning. Another broad set of machine learning methods is described by the term “reinforcement learning.” Reinforcement learning involves an agent interacting with its environment. As the model interacts with the environment, it learns to make better decisions over time based on a reward function. Newer AI applications will need to operate in increasingly dynamic environments, and react to changes in those environments, which makes reinforcement learning a useful technique. Reinforcement learning has several attributes that make it a distinctly different engineering problem than supervised learning. Reinforcement learning relies on simulation and distributed training to rapidly examine how different model parameters could affect the performance of a model in different scenarios. Ray is an open source project for distributed applications. Although Ray was designed with reinforcement learning in mind, the potential use cases go beyond machine learning, and could be as influential and broadly applicable as distributed systems projects like Apache Spark or Kubernetes. Ray is a project from the Berkeley RISE Lab, the same place that gave rise to Spark, Mesos, and Alluxio. The RISE Lab is led by Ion Stoica, a professor of computer science at Berkeley. He is also the co-founder of Anyscale, a company started to commercialize Ray by offering tools and services for enterprises looking to adopt Ray. Ion Stoica returns to the show to discuss reinforcement learning, distributed computing, and the Ray project. If you enjoy the show, you can find all of our past episodes about machine learning, data, and the RISE Lab by going to SoftwareDaily.com and searching for the technologies or companies you are curious about . And if there is a subject that you want to hear covered, feel free to leave a comment on the episode, or send us a tweet @software_daily. Sponsorship inquiries: sponsor@softwareengineeringdaily.com The post Anyscale with Ion Stoica appeared first on Software Engineering Daily.

<p>Machine learning applications are widely deployed across the software industry. </p> <p>Most of these applications used supervised learning, a process in which labeled data sets are used to find correlations between the labels and the trends in that underlying data. But supervised learning is only one application of machine learning. Another broad set of machine learning methods is described by the term “reinforcement learning.”</p> <p>Reinforcement learning involves an agent interacting with its environment. As the model interacts with the environment, it learns to make better decisions over time based on a reward function. Newer AI applications will need to operate in increasingly dynamic environments, and react to changes in those environments, which makes reinforcement learning a useful technique.</p> <p>Reinforcement learning has several attributes that make it a distinctly different engineering problem than supervised learning. Reinforcement learning relies on simulation and distributed training to rapidly examine how different model parameters could affect the performance of a model in different scenarios.</p> <p>Ray is an open source project for distributed applications. Although Ray was designed with reinforcement learning in mind, the potential use cases go beyond machine learning, and could be as influential and broadly applicable as distributed systems projects like Apache Spark or Kubernetes. Ray is a project from the Berkeley RISE Lab, the same place that gave rise to Spark, Mesos, and Alluxio.</p> <p>The RISE Lab is led by Ion Stoica, a professor of computer science at Berkeley. He is also the co-founder of Anyscale, a company started to commercialize Ray by offering tools and services for enterprises looking to adopt Ray. Ion Stoica returns to the show to discuss reinforcement learning, distributed computing, and the Ray project.</p> <p>If you enjoy the show, you can find all of our past episodes about machine learning, data, and the RISE Lab by going to <a href="https://www.softwaredaily.com/" rel="nofollow" target="_blank">SoftwareDaily.com</a> and searching for the technologies or companies you are curious about . And if there is a subject that you want to hear covered, feel free to leave a comment on the episode, or send us a tweet @software_daily.</p> <p>Sponsorship inquiries: <a href="mailto:sponsor@softwareengineeringdaily.com">sponsor@softwareengineeringdaily.com</a></p> <p>The post <a href="https://softwareengineeringdaily.com/2020/02/13/anyscale-with-ion-stoica/" rel="nofollow" target="_blank">Anyscale with Ion Stoica</a> appeared first on <a href="https://softwareengineeringdaily.com" rel="nofollow" target="_blank">Software Engineering Daily</a>.</p>