CentML

Big news from CentML: Meta's Llama 3.3 70b is now live on our platform! https://lnkd.in/g4gmt-EZ In significant improvement over the previous generations, Llama 3.3 packs the same performance as Llama 3.1 with 405 Billion parameters in fraction of the cost and compute footprint with a smaller form factor. With CentML’s LLM serving platform, you can deploy Llama 3.3 on a range of AI hardware - NVIDIA H200s, Google Cloud Trillium, AMD MI300X - all in one place. Ready to unlock Llama 3.3? Start deploying today: https://lnkd.in/g4gmt-EZ #LLM #AIInnovation #CentML #Llama3.3

Big news from CentML: Meta's Llama 3.3 70b is now live on our platform! https://lnkd.in/g4gmt-EZ In significant improvement over the previous generations, Llama 3.3 packs the same performance as Llama 3.1 with 405 Billion parameters in fraction of the cost and compute footprint with a smaller form factor. With CentML’s LLM serving platform, you can deploy Llama 3.3 on a range of AI hardware - NVIDIA H200s, Google Cloud Trillium, AMD MI300X - all in one place. Ready to unlock Llama 3.3? Start deploying today: https://lnkd.in/g4gmt-EZ #LLM #AIInnovation #CentML #Llama3.3

🚀 The wait is over — CentML's next-gen platform is HERE! Harness the power of NVIDIA H200, Google TPU v6e, and Amazon Web Services (AWS) Inferentia2 all in one place 😍 (AMD MI300X coming soon!) But we didn’t stop there. In our experience working with customers, we found that the LLM itself is often not enough to build a fully AI-powered application. That’s why we’ve introduced 1-click deployable RAG on CentML Platform — to streamline development and take the complexity out of integration. CentML now brings you access to the world’s most powerful AI chips, combined with the best prices in the market. Start today: https://lnkd.in/g4gmt-EZ #artificialintelligence #AIengineering #machinelearning #MLengineering

Headed to AWS re:Invent this week? The CentML team is on-site, ready to answer your questions about deploying superior LLMs on the world's most powerful AI chips 🦾 Let's connect! 👋 https://centml.ai/contact/ 💚 Get $10 in free credits on the CentML Platform: https://app.centml.com/ ⚙️ Learn more about the platform: https://lnkd.in/gA--tkEH 💸 Review our flexible deployment pricing: https://centml.ai/pricing/ Amazon Web Services (AWS) #reinvent #AWSreinvent #awsreinvent2024 #AIengineering

Introducing a one-stop shop for deploying superior LLMs on the world's most powerful AI chips! Harness the power of NVIDIA H200s, Google Cloud Trillium, and AMD MI300X, all in one place 🤯 CentML’s next-gen platform provides access to the world’s best chips, running GenAI and LLMs like Llama 3.1, Mixtral, QWen, and more with ease. Launching in December 🚀 Sign up for early access today! https://lnkd.in/gZ3n-5VC #CyberMonday #AI 

Multi-cloud deployment can be incredibly complex, but there is a better way 🕸️✨ With our open-source tool ‘ECR Anywhere’ you can help ensure secure, seamless deployments → https://lnkd.in/gdfawBxG Native registries work well on their own turf, but create complexity in cross-cloud environments. Here’s how ECR Anywhere transforms multi-cloud management: → Streamlines hosting Docker images across cloud environments → Simplifies credential management for Kubernetes clusters outside AWS → Maintains the security and convenience of ECR without compromising flexibility Whether you're deploying AI models, scaling enterprise apps, or testing prototypes, ECR Anywhere supports frictionless cross-cloud operations. 💡 Learn more: https://lnkd.in/gdfawBxG  🚀 Get started: https://lnkd.in/gT4gEzUC #DevOps #CloudComputing #AIdevelopment #MLdevelopment #OpenSource

One of the biggest barriers to AI adoption is the sheer cost of deployment 💰 That’s why we provide transparent, flexible pricing, and millions of free tokens for CentML Platform users 🦾 www.centml.ai/pricing/ With rapid, affordable deployment of superior GenAI, developers and enterprises alike can build, deploy, and scale powerful AI models — without hidden fees or added complexity. 👩💻 Developer-friendly pricing: → Pay-as-you-go model ensures you only pay per token or minute used → No rate or daily usage limits → On-demand dedicated endpoints allow for seamless scaling 📊 Enterprise solutions: → Custom deployments and pricing to support large-scale AI needs → Guaranteed uptime, unlimited deployments, and 24/7 support Start your first deployment on the CentML Platform with $10 in free credits. That's enough to run 4 million tokens on Llama 3.1 405B! 🦙🦾 #AI #MachineLearning #GenAI #CentML #AIforAll

Great experience talking about AI, LLMs, GPUs, and many related topics with Tim Scarfe. Really enjoyed it! Always happy to share both our industrial experience at CentML and academic experience at the University of Toronto. #genai #llm #optimization #gpu #ml #centml

We are wasting a tonne of GPU cycles on typical ML workloads, especially in the age of DIY LLMs. Many choose DIY because they don't want to send their company data to OpenAI and have specific architectural requirements. Because of the current cloud model, we see a proliferation of individual cloud subscriptions with GPUs idling all over the place (GPUs are not yet widely virtualised in the way CPUs are on the cloud). Eventually we will graduate to a centralised/virtualised model with dynamic/active compiler optimisations for specific ML workloads, this is what I learned from Gennady Pekhimenko from CentML

We're expanding our Sales team and hiring our first SDR here @ CentML! 📈 🤝 💰 Are you interested in how system architecture across GPU, networking, CPU and IO relate to brand new generative AI capabilities? Come join our team, and bring your experience and interests to help us sell the next generation of inference and training frameworks to redefine the world. Job Posting - https://lnkd.in/gX7fr6Pg 'Image generated by DALL·E' 🤖