Yousef Khalidi

💥JUST IN: NVIDIA + Dell Technologies + U.S. Department of Energy (DOE) + Berkeley Lab announce Doudna Supercomputer named after Nobel laureate Jennifer Doudna (pictured on the left)! This system based on our next-generation Vera Rubin (CPU + GPU) architecture will serve a community of 11,000+ scientists and researches in the DoE ecosystem. Official press release: https://lnkd.in/eaPz-ctc Dion Harris NVIDIA Data Center

1,674

24 Comments

I'm excited to share that our partnership with Microsoft will now bring you full support for Azure ML services on VMware Cloud Foundation. While our customers and partners continue to embrace #PrivateAI, they have also asked for broadening support for #HybridAI alternatives that give them choice of best-of-breed in the public cloud and on-premises or edge. We've been listening, and that's what makes today's news especially exciting. Check out the blog post below for more information and a link to our solution architecture authored by Fabiano Teixeira. https://lnkd.in/gEKzpiMA

614

8 Comments

GraphRAG has emerged as a frontrunner in building question-answering systems that enable organizations to extract relevant insights from their interconnected data. In this blog, we demonstrate how to leverage LangChain and Spanner Graph to build powerful GraphRAG applications. Worth a weekend read...!

192

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

Today, Microsoft joined Google's developed Agent2Agent (A2A) framework ecosystem to contribute to, and advance the protocol, including interoperability with Azure AI Foundry and Copilot Studio. This is good news for the entire ecosystem! Next to MSFT, we have already welcomed over 50 major Tech players who, together, are shaping the world of AI Agents. Onwards and upwards 🚀

562

13 Comments

Data can be messy. Engineering it does not have to be. With today’s acquisition of Osmos, Microsoft is accelerating autonomous data engineering in Fabric so soon AI agents can do more of the heavy lifting alongside people. The goal is faster insights, lower operational drag, and data that is ready to work when you are. Full details on the Official Microsoft Blog: https://lnkd.in/ggMw9JaG #Microsoft #MicrosoftFabric #PowerBI

847

30 Comments

Pragmatic Engineer did a great deep dive on "Building MCP servers in the real world": https://lnkd.in/ghBu9fwV Great insights on the growing use of *internal* MCP servers, like for documentation, data analytics, debugging, issue triaging, etc. That's why MCP auth and networking is so important. I highly recommend the Pragmatic Engineer newsletter -- I finally upgraded to a paid subscription today, reimbursed by Microsoft in PD budget.

162

1 Comment

We just shipped an update to Agentic Retrieval in Azure AI Search that boosts results quality when grounding agents on external data for elaborate scenarios where a single search against a single data source would not be enough. We're expanding Agentic Retrieval with the ability to target multiple indexes in a single operation. This is not just a fan-out query, the query planner uses the information it has about each source of data to intelligently decompose retrieval tasks into separate subtasks, issues the right queries to the right indexes, and then composes a unified result ready to be sent to the language model backing an agent. The system is built and evaluated from the ground up with consideration of steerability. Developers can provide descriptions and instructions for each data source, as well as overall instructions for the retrieval agent that oversees the operation. We also added the option for answer synthesis in the same retrieval call. When answer synthesis is enabled, we not only return the raw grounding information but also an actual answer to the question or task based on the various results from the knowledge sources the system selected. With this capability, a single call to the /retrieve API replaces what would have been extensive context engineering work to get routing and query planning right, several calls to a language model for query decomposition, several calls to separate search indexes to actually retrieve data, and subsequent work to stitch results together. More in Matthew Gotteiner's blog post: https://lnkd.in/ggyF42Jw

577

24 Comments

GraphRAG can be a more effective way to perform RAG (Retrieval Augmented Generation) in Generative AI. Below is a good intro to how to do GraphRAG, using LangChain and Spanner Graph. https://lnkd.in/e79KXN5k

46

Still confused between MCP, A2A, and Functional Calling? These three approaches are reshaping how LLMs and agents operate at scale, but each one solves a very different problem. Let’s break it down 👇 1. Model-Context Protocol (MCP) MCP manages dynamic context windows across enterprise LLM infrastructure. It retrieves prior memory, history, and metadata to personalize every response without hardcoding logic into prompts. 2. Agent-to-Agent Protocol (A2A) In A2A, agents talk to each other directly using shared protocols. Think of it as collaborative problem-solving between specialized local agents, especially useful in multi-agent ecosystems. 3. Functional Calling Here, an LLM acts like a planner. It decomposes a task (e.g., comparing market caps), generates a Directed Acyclic Graph (DAG) of subtasks, and then sends each task to appropriate tools (math, search, APIs) for execution. Conclusion - MCP is about managing context, - A2A is about agent collaboration, - Functional Calling is about task execution. Mastering all three is key to building powerful, production-ready AI systems. Save this visual for your next AI project!

357

52 Comments

From Idea to Impact: How Panda AI (powered by Moveworks) Is Redefining Productivity at Palo Alto Networks 🐼 When generative AI exploded into the mainstream, Palo Alto Networks didn’t just experiment—they operationalized. Led by CIO Meerah Rajavel, the company crowdsourced use cases across every department, but only pursued the ones with the clearest ROI. The standout? Panda AI—an internal agent that now autonomously resolves 60% of all IT, HR, and Finance requests from employees. This wasn’t just about automation. It was about rethinking support: 📞 Phone lines? Gone. 💬 Slack channels? Gone. 🧾 Online portals? Gone. Every employee starts with Panda. Only if the agent can’t complete the task does it escalate to a human. The payoff? Huge cost savings. Fewer outsourced agents. And a re-skilled internal team now focused on fine-tuning AI accuracy, not repetitive ticket handling. It’s a paradigm shift in enterprise support—from information retrieval to task completion. Panda is more than a chatbot—it’s a lesson in focus, intentionality, and change management. The implications? ✅ AI won’t replace your teams—but it will reshape them. ✅ The biggest wins come when AI handles actions, not just answers. ✅ Adoption starts not with tech, but with trust and a willingness to rewire how work gets done. This is what the future of AI at work actually looks like. Thank you Meerah Rajavel and Palo Alto Networks team for your partnership! 🔗 Link in comments.

83

4 Comments