Amit Zavery

Strong look at how our differentiated portfolio and customer partnerships are driving Google Cloud's growth story coming out of the Goldman Sachs' Communacopia conference yesterday: https://lnkd.in/gktdMFy4 As Thomas says, "We've spent years building advanced AI technology of our own — chips, systems, tools, agents — we made those bets very early. Much of the work that you see today has been underway for many, many years. We're not just reselling third-party AI technology. So why we're winning is because we see this deep product differentiation now being adopted by customers and (broadening) our addressable market. In turn, that's leading us to grow revenue and operating income."

242

3 Comments

We're excited to announce new capabilities in Google Cloud #VertexAI Training designed to simplify and accelerate large-scale model development. Key highlights from our latest update: 🔹 Flexible, Self-Healing Infrastructure: Leverage fully managed, resilient Slurm environments with automated failure detection and performance-optimized checkpointing. 🔹 Cost-Effective Scheduling: Utilize Dynamic Workload Scheduler (DWS) for fixed future reservations or flexible on-demand capacity. 🔹 Integrated Frameworks: Access optimized recipes for the full model lifecycle (including SFT and DPO) and seamless integration with NVIDIA NeMo. Whether you're fine-tuning standard models or training massive custom foundational models from scratch, these new features are built to get you to production faster. Read the full announcement to see how organizations like Salesforce and AI Singapore are already leveraging these tools to elevate their models. 👇 https://lnkd.in/gVVq3WCw

185

2 Comments

I thoroughly enjoyed moderating a panel on quantum computing for executive leaders at SIBF this week in Palo Alto with Amit Bhattacharyya (UC Berkeley), Charles Chung (IBM), and Rajesh Ramanujam (Applied Ventures). We talked about a broad future "Beyond AI" and explore the various ways quantum is advancing. A few key takeaways: 1 - Chuck shared that IBM anticipates demonstrating "Quantum Advantage" next year, with a fault tolerant computer by 2029. 2 - Rajesh shared how companies concerned about data security should explore "Post Quantum Cryptography" 3 - Amit expounded upon how AI is accelerating quantum research, and that no matter how helpful AI may be today, students still must develop core critical thinking skills.

72

3 Comments

AI chatbots are super cool but how does one make that serve real-world enterprise tasks? OCI AI Agent Platform, a next frontier in Generative AI, automates complex real-world enterprise workflows. In the latest first principles blog, Egor Pushkin and I discuss how OCI AI Agent Platform works: https://lnkd.in/gN9kC6mC

410

6 Comments

Standing up an AI workload shouldn’t take months, but a steep learning curve and a rapidly evolving stack often stretch production LLM deployments far beyond plan. We recently launched OCI AI Accelerator packs to address the most common blockers we see in enterprise AI. These packs reduce the time required to deploy Generative AI applications, while helping enterprises reduce the learning curve and stack complexity that typically slows production and deployment. Learn more: https://lnkd.in/emiTNsRF

189

3 Comments

While LLMs in the cloud have enamored consumers globally, what will scale in enterprise and physical AI industries globally, will be SLMs (Small Language Models) and on device - cost and power efficient, safe, secure, and with no need for the cloud. The next big wave of AI adoption will be in the physical world. Read the article at Financial Express (India) https://lnkd.in/gpZnT4-y

90

3 Comments

📣 kgateway now supports Kubernetes Gateway API Inference Extension! 📣 AI workloads backed by GPUs running on Kubernetes are NOT like typical webapps. You need to solve the GPU multiplexing problem, but you also need to solve the request load-balancing problem. AI inference requests have unique characteristics that make conventional load-balancing techniques less effective. kgateway helps with this problem. Make smarter decisions based on real-time metrics such as GPU cache utilization, LLM queue depths, model names/fine-tuned adapters loaded, etc, etc. 🔎 Read more in my latest blog: https://lnkd.in/gTsna8wj 🔎 Checkout the Inference Extension OSS community: https://lnkd.in/g6gqyBju 🔎 Get started with kgateway: https://lnkd.in/gqCVE29a #ai #inference #kubernetes #networking #gatewayapi #apigateway #kgateway #cncf #llm

43

3 Comments

Our new article "Building and deploying prebuilt domain agents in watsonx Orchestrate" co-authored with watsonx Orchestrate Product Managers (Kayla Regulski and Srinidhi Kunigal) in the Agentic AI series is now live on IBM Developer: https://lnkd.in/dxqiBrNe IBM watsonx Orchestrate Agent Catalog provides a diverse selection of prebuilt domain agents across functions like HR, sales, customer care, finance, IT, and more. As of today, the catalog features 232 prebuilt domain agents and 831 prebuilt tools, each capable of performing specific actions, retrieving information, or complex reasoning. This rich ecosystem is expanding rapidly, delivering ongoing access to new capabilities and enterprise applications. In this guided walkthrough, you will explore the robust library of prebuilt domain agents in IBM watsonx Orchestrate, showcasing its Agent Catalog and diving deeper into pre-built domain agents for HR agents with SAP SuccessFactors and Sales Prospecting agents with Salesforce, Dun & Bradstreet, Seismic, and Outlook. The value of prebuilt domain agents lies in their ability to combine structured automation with intelligent reasoning, enabling organizations to orchestrate complex, multi-step processes without coding. They reduce manual effort, improve accuracy, and accelerate solution delivery while maintaining transparency and control. Let’s build together, with watsonx. 🏗️ #AgenticAI #watsonx Suzanne Livingston | Vinicius Maidana | Mithun Katti | Matt Sanchez | Dawn Herndon | Bill Higgins | Sabine Holl | Sampath Dechu | Ela Dixit | Goura Mohanty | Peter A. Keller | Allen Chan | Ryan Rivera | Bostjan Kos | Elena Lowery | Asli Erbas | Hande Akdede Erbay | Zaidoun Arbad | Grant Millar | Ben Nicholls | Frank Ketelaars | Jyotsna K Narayanan | Merve UNUVAR | Vinod Khader | Michelle Corbin | Bindu Umesh, PMP® | Giulio S.

108

Excited for today's launch of Vertex AI Training, our new managed service on Google Cloud that simplifies and accelerates the path to developing large, highly differentiated models. I spoke with VentureBeat about this new offering, read more at the link below! These new features, designed for customization and scale, combine the best of Google Cloud AI infrastructure with sophisticated management tools. This includes a fully managed and resilient Slurm environment based on Cluster Director, prebuilt data science tooling, and optimized recipes integrated with frameworks like NVIDIA NeMo — for specialized, large-scale model building. https://lnkd.in/gJc82QQF

238

2 Comments