We’re thrilled to share that our co-founder Robert Nishihara has been named to Fast Company’s AI20, recognizing leaders shaping the future of AI! 🚀 Robert’s AI story started when he was doing AI research at UC Berkeley under Ion Stoica and Michael Jordan. Despite being researchers, Robert and co-founder Philipp Moritz were spending all of their time on distributed systems challenges like managing clusters. So they created Ray, an open-source, Python-native compute engine that simplifies scaling AI – from laptops to the cloud – with just a few lines of code. Ray is the foundational AI Compute Engine for over 10,000+ companies enabling them to scale and create better models with leading performance and efficiency: ✅ Amazon reduced costs by $120M per year by switching to Ray ✅ Pinterest reduced model training time by 6x ✅ Niantic cut lines of code by 85% Even OpenAI trained GPT-4 using Ray. Anyscale, founded in 2019, builds on Ray with a unified AI Platform and an optimized version of Ray called RayTurbo. Customers like Canva, Instacart, and RunwayML are building their AI platforms powered by Anyscale. ✅Instacart trains models with 100x more data ✅ Canva slashed cloud costs by 50% “We’re at the outset of tremendous value being created by AI,” Nishihara says. “To realize that value, much of the hard work ahead involves taking existing (and future) capabilities, making them incredibly reliable in the real world, and building out the underlying hardware and software infrastructure to enable them throughout every industry.” Read how Robert Nishihara and the team behind Ray are shaping AI’s future: https://lnkd.in/gYJv2qJ2
Anyscale
Software Development
San Francisco, California 37,011 followers
Scalable compute for AI and Python
About us
Scalable compute for AI and Python Anyscale enables developers of all skill levels to easily build applications that run at any scale, from a laptop to a data center.
- Website
-
https://anyscale.com
External link for Anyscale
- Industry
- Software Development
- Company size
- 51-200 employees
- Headquarters
- San Francisco, California
- Type
- Privately Held
- Founded
- 2019
Products
![Click here to view Anyscale]()
Anyscale
AIOps Platforms
The Anyscale Platform offers key advantages over Ray open source. It provides a seamless user experience for developers and AI teams to speed development, and deploy AI/ML workloads at scale. Companies using Anyscale benefit from rapid time-to-market and faster iterations across the entire AI lifecycle.
Locations
-
Primary
Get directions
San Francisco, California 94105, US
Employees at Anyscale
Updates
-
Anyscale reposted this
Generative models for 3D have gotten less attention but will be very very useful for design. In the same way that image generation models make it far easier for non-artists to make visual designs, these tools will enable us to easily design and create 3D objects, architecture. Very cool to see projects like Bernini built with Ray. Autodesk shared how they preprocess data and build models at Ray Summit 2024. https://lnkd.in/eTWvgSv6
Research Project Bernini is an Autodesk experimental generative AI that can quickly generate multiple functional 3D shapes from a variety of inputs, including text, 2D images, or a voxel. Project Bernini is at the leading edge of generative AI, and its models will become increasingly useful and compelling when trained on larger, higher-quality professional datasets. Stay tuned for future updates as we continue to develop cutting-edge AI that helps our customers design and make anything: https://bit.ly/4bqcVU3.
-
This was one of the most popular talks at Ray Summit 2024.
Nvidia's video data curation pipeline runs inference with 7+ models (and many data curation pipelines use far more). "Data curation" is a very very important part of training. There are two main parts (both of which are use case dependent) - *Filtering:* remove low quality or uninteresting data - *Annotation:* extract structured information, compute embeddings, categorize the type of video, generate text descriptions (in the case of video), etc (very problem dependent) There are a ton of technical challenges here and lots of heterogeneity. - In the inference stages, different models have different throughput, which needs to be balanced. - Scale is massive: e.g., 100s of petabytes, 1000s of H100s, millions of hours of video processed per day. - Failures are common. - Heterogeneity requires streaming execution (otherwise you leave your most important resources under-utilized). The full video on NVIDIA's video data curation pipeline from Ray Summit is here: https://lnkd.in/gRsZwPpm
How NVIDIA is Advancing Video Curation with Generative AI | Ray Summit 2024
https://www.youtube.com/
-
Anyscale reposted this
This vision is spot on. Today, the *main* way that companies get value from data is through SQL queries (and relatively simple analytics). SQL is amazing and isn't going anywhere (and will continue to grow), but it's also limited. You aren't going to use SQL to get insights from your video data or from random PDFs lying around or other unstructured multimodal data. You aren't going to use SQL to deeply reason and draw conclusions about your customer base from all the meeting notes and recorded calls that your team has produced. That is going to be done with AI. The process of "getting value from data" will shift from being SQL-centric to being AI-centric.
Today, Box announced its Q3 results, with revenue coming in at $276M and record non-gaap op margin of 29%, and we guided to approximately $1.09B in revenue and op margin of 28% for the full year. These results are driven by the continued growth in demand for Box AI, and the need for enterprises to transform how they work with their content. The vast majority of an enterprise’s data is unstructured data, with enterprise content being the most significant portion of this. Some of the world’s most important IP lives inside of this content — our contracts, financial documents, digital assets, movie scripts, R&D files, product specs, HR documents, and more. This is the content that makes blockbuster films, gets a new product to market, delivers a breakthrough new drug, launches a killer new advertising campaign, or closes the books at the end of the quarter. However, the majority of this content has tremendous amounts of untapped value for most organizations. For all of our *structured* data that lives inside of a database, we can query this information easily, summarize it, analyze it, pull out insights, and more. But for all of our unstructured data, this has been nearly impossible. In fact, in many cases the more content we have the harder it is to work with and less valuable it becomes. In this new era of AI, this all is reversed. What if you could ask all your content any question you want, or automate any of the workflow instantly. All new ways of working with information become possible: “What’s our best performing product line?”, “How many contracts have risky terms in them?”, “Which clients do we have promotional rights to?”, “Show me all my open invoices”. All of a sudden, the more information we have the more valuable it becomes. We can innovate more and accelerate progress. Now, enterprises are still in the earliest stages of beginning to leverage AI on their unstructured data, but with Box AI we’re building the platform to make this easy, secure, and scalable for any use-case. And we’re doing so with an open approach, to bring the power of any AI model into Box AI so customers can leverage whatever works best. That’s the power of intelligent content management from Box. It’s an incredibly exciting time to be in tech, and I’m insanely proud of the execution of the team — and deeply appreciative to all our customers and partners for helping us get here. And most importantly… we’re just getting started.
-
At #reinvent2024? Stop by booth #1920 to learn how RayTurbo drives scalable, high-performance LLM applications. 🔥 Discover how Anyscale is transforming AI development with advanced solutions for efficiency and growth. 🚀 See you there! #AWSreInvent #LLM #RayTurbo #Ray #Anyscale
-
Exciting news! 🎉 Anyscale is joining Amazon Web Services (AWS) for the AWS GenAI Live show at re:Invent! When: Tuesday, December 3rd, 2:20-2:40pm PT Where: Streaming LIVE on AWS Partner Network YouTube https://lnkd.in/gnAJCYgu Our team will discuss the AI Complexity wall and how Ray, the AI Compute Engine, helps scale AI for multimodal data processing, model training, and model serving workloads in GenAI. Register now: https://lnkd.in/gKZqa4N3 Watch LIVE: https://lnkd.in/gnAJCYgu Join us as we explore what’s next for AI.💡 #AWSreInvent #AWSGenAILive #generativeAI #AWSPartners #Ray #Anyscale
-
-
Will you be at AWS re:Invent 2024? Join MongoDB, Anyscale, Cohere, and Fireworks AI for an exclusive panel on how to leverage the best tools to create scalable, high-performance #genAI applications. - Date: Wednesday December 4, 2024 - Time: 1pm - Location: Bollinger @ Wynn Las Vegas Don't miss your chance to hear insights from the top leaders driving innovation in the AI space. See you there!
-
-
Anyscale reposted this
Patrick Ames from AWS has more experience than just about anyone with managing thousands of Ray clusters and Spark clusters and processing exabytes of data. There are tremendous operational and performance challenges. - managing thousands of clusters - scaling to many petabytes or exabytes - latency and cost The original blog post is here: https://lnkd.in/g-pJhFei https://lnkd.in/gQsCrXPH
Is Apache Spark Too Costly? An AWS Engineer Tells His Story
https://www.youtube.com/
-
Are you headed to #reinvent2024? Join us on Dec. 4th for a hands-on workshop with Amazon Web Services (AWS) and learn how to scale generative AI workloads! This session will cover: 🔹 Building LLM apps with RayTurbo, Anyscale’s optimized Ray runtime. 🔹 Advanced scaling: autoscaling, CPU/GPU integration, replica compaction. 🔹 Fine-tuning LLMs with LLM-Forge. 🔹 Debugging and profiling with Anyscale observability tools. Seats are limited—register here 👇 https://lnkd.in/gqRYFPz6
Generative AI Workloads with Anyscale and AWS
learn.anyscale.com
-
Will you be at Amazon Web Services (AWS) re:Invent 2024 next month? Join us, MongoDB, Cohere and Fireworks AI on Dec 4 at 1 PM (Wynn Las Vegas, Bollinger room) for a panel on building scalable, high-performance #genAI apps. 🙌 Hear from top AI leaders and learn how to level up your AI stack. 🚀 Don’t miss it! #reInvent2024 #Anyscale #AI
-