Lakshmanan Sethu Sankaranarayanan

Happy to share my new blog: https://lnkd.in/g4zHR49p SageMaker Introduces S3-Based Shared Storage! written in collaboration with Hari Ramesh Zachary Mitchell Anchit Gupta and Anagha Barve 🔹 𝗪𝗵𝗮𝘁'𝘀 𝗡𝗲𝘄: Amazon SageMaker now offers Amazon S3-based shared storage as the default project file storage option for SageMaker Unified Studio projects, replacing AWS CodeCommit. 💡 𝗪𝗵𝗮𝘁 𝗧𝗵𝗶𝘀 𝗠𝗲𝗮𝗻𝘀 𝗙𝗼𝗿 𝗬𝗼𝘂:  • No more complex Git operations for basic file sharing  • Seamless access across various SageMaker tools  • Clear separation between personal and shared workspaces  • Global availability in all AWS Regions where SageMaker is supported 🌟 𝗞𝗲𝘆 𝗕𝗲𝗻𝗲𝗳𝗶𝘁𝘀:  • Simplified collaboration - Share files directly with team members  • Universal access - Consistent file experience across JupyterLab, Query Editor, and Visual ETL  • Personal workspace separation with EBS volumes  • Optional Git integration still available for version control  • Immediate file visibility for all project members Ready to enhance your team's collaboration in SageMaker projects? Explore the new S3-based shared storage in SageMaker today! 𝗕𝗹𝗼𝗴: https://lnkd.in/g6DWYpMx

43

At Build, we are announcing a series of exciting updates to AI Search. First: Agentic Retrival with Azure AI Search. Let AI Search handle all retreival complexity on behalf of yourself with query decomposition, multi query retrieval and more.. https://lnkd.in/gK5S_Z6F #build #ai #search #rag #llm #openai

194

4 Comments

After listening to Satya talk about the pace of AI-first feature development across Microsoft suite of products in our FY25 earnings, I thought I would take a moment to share one AI-first feature per every major product that I touch quite frequently, and is integral to our AI diffusion capabilities. 1. Copilot: Analyst is amazing if you haven't tried it. I am a huge data geek - and being able to just link to massive excel files and let Analyst write code, think through the data, and answer compound questions is pretty sweet. 2. Copilot Studio: Copilot Studio Agent Analytics got a revamp a couple of months ago, and it's really useful. I can review if my agents are using the right tools, if they're processing the right data in a high-quality way, and also nudge the agent instructions to get a higher "effectiveness" score. 3. Azure AI Search: AI Search is SO underrated. It has the potential to be your unified lakehouse equivalent for your entire organization's semantic knowledge. And we recently announced agentic retrieval in Search, meaning you don't have to fuss over the right search query - the retrieval agent will auto-decompose your asks and aggregate the relevant responses. 4. Azure AI Foundry: We have o3-deep-research as an API in Foundry now - meaning you can build deep research experiences across your proprietary banking, wealth, healthcare, retail data without having to share data publicly. Pull that into our Agent Service, and you can have an agent that does deep research, connects to your product catalog, and starts actioning recommendations. 5. Fabric: Fabric now has CosmosDB as a database offering, and Cosmos is how both ChatGPT and Copilot manage memories and conversation threads. I think unifying semantic data with structured data is a huge differentiator, and having both live in one pane of glass makes life easier for data engineers. 6. Azure API Management (APIM): APIM is the AI gateway for your agents, for both the brains (models) and the limbs (tools). With MCP support native in APIM, and ability to manage and scale across multiple model providers, APIM is a really integral node in the agentic orchestration. I am curious if folks are leveraging other products across the ecosystem that have had a cool new release recently!

33

2 Comments

Oracle just laid off nearly 10% of its workforce in India – thousands of people across cloud, engineering, and ops roles. At the same time, Oracle signed a $300B deal with OpenAI to embed AI deeply into its enterprise applications. 👉 What does this mean? Enterprise software is changing. The skill sets companies need are shifting from manual ops & redundant engineering → towards AI-driven automation, orchestration, and system design. ⚙️ In reality, enterprise AI is not just about training ML models. It’s about: -RAG to ground LLMs on enterprise data -Summarization & Extraction to streamline workflows -Automation pipelines to replace repetitive human tasks -AI Engineering to stitch all of this into production systems 💡 And here’s the truth: It’s rarely “just RAG” or “just chat with docs.” In industry, you engineer combinations of RAG, summarization, extraction, and automation — based on user requirements and enterprise constraints (latency, cost, compliance, explainability). 🚀 That’s why our AI Bootcamp was built differently. Not just core ML. But overall automation + AI engineering for enterprises. The Oracle layoffs are not just cost cuts. They’re a signal: Enterprise software companies are re-allocating resources → fewer “support engineers,” more AI engineers. If you’re working in enterprise software today, ask yourself: ➡️ Do you know how to design AI + automation into your stack? ➡️ Can you engineer AI systems that actually fit enterprise workflows? That’s exactly what we teach in the AI Engineering Course. 👉 Sign up here: https://lnkd.in/g89QQDrb Shashank Mishra 🇮🇳 SHAILJA MISHRA🟢 Rahul Shukla Sahil Choudhary Aman Kumar Grow Data Skills #learning #ai #tech #layoffs

284

1 Comment