Mónica Lora

🎙️ How does Confluent enable cost management in the cloud? 🤝 Italo Nesi and John Dee reconvene for the fifth episode in our capabilities podcast series, to explore Confluent's consumption-based pricing model, which allows businesses to accurately forecast expenses based on their application usage. 🖥️ Watch episode #5 now: https://lnkd.in/eeMMP_Xq Paul Archer Joanna Grimmond Chris Gallagher #Confluent #CostManagement #CloudCosts

5

Redshift Serverless just became better with the GA of AI scaling and optimizations.

64

1 Kommentar

the upcoming week, I’ll participate in a panel discussion on Open Source and AI, how it transforms businesses and IT systems, basically on the fact, that transparency of AI-driven systems is extremely important! Some of us call it Trustworthy AI https://lnkd.in/eY3ZHYEg

21

𝗪𝗵𝗮𝘁'𝘀 𝗵𝗼𝘁 𝗶𝗻 𝗔𝗜 - 𝘄𝗵𝗮𝘁 𝗰𝗮𝗺𝗲 𝗳𝗿𝗼𝗺 𝗜𝗕𝗠, 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗮𝗻𝗱 𝗚𝗼𝗼𝗴𝗹𝗲 𝗯𝗶𝗴 𝗮𝗻𝗻𝘂𝗮𝗹 𝗰𝗼𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲𝘀 𝗶𝗻 𝗹𝗮𝘀𝘁 𝟮 𝘄𝗲𝗲𝗸𝘀 The past two weeks have been full of innovation in the AI industry, with major announcements from IBM, Microsoft, and Google during their annual conferences. The highlights below reflect my own observations and common themes across vendors. I can see few emerging trends, showcasing the industry's focus on: ✅ Open-source Innovation ✅ AI Agents / Assistants ✅ Small Language Models ✅ Custom AI Chips 1️⃣ 𝗜𝗕𝗠 𝗧𝗵𝗶𝗻𝗸 : 𝗢𝗽𝗲𝗻 𝗦𝗼𝘂𝗿𝗰𝗲 𝗙𝗼𝗰𝘂𝘀 𝗮𝗻𝗱 𝗡𝗲𝘅𝘁-𝗚𝗲𝗻 𝗔𝗜 𝗔𝘀𝘀𝗶𝘀𝘁𝗮𝗻𝘁𝘀 https://shorturl.at/fk7Yf 🔹 IBM released its Granite family of foundation models into open-source. Developers can now use and modify these models under Apache 2.0 license, 🔹 InstructLab Model Alignment: Partnering with Red Hat, IBM introduced InstructLab, a groundbreaking model alignment technique, integrating community contributions into LLMs, streamlining development process. 🔹 AI Assistant Builder unveiled within Watsonx Orchestrate platform simplifies the creation of custom AI assistants tailored to specific needs, accelerating enterprise AI adoption. 🔹 Strategic Partnerships: IBM announced extended partnerships with industry leaders like AWS, Adobe, Meta, Microsoft, Salesforce, SAP, Palo Alto and others. This includes Allam, an LLM from SDAIA (Saudi Data & AI Authority) which is already available on the Watsonx.ai platform. 2️⃣ 𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗕𝘂𝗶𝗹𝗱: 𝗖𝗼𝗽𝗶𝗹𝗼𝘁 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀 𝗮𝗻𝗱 e𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗦𝗟𝗠𝘀 https://shorturl.at/ZFXC2 🔹 Copilot AI Agents and GPT-4o: Microsoft's focus revolved around Copilot Studio and their partnership with OpenAI. GPT-4o is at core of multi-modal agents spanning text, speech, and vision. 🔹 Small Phi-3 Models: Microsoft introduced the Phi-3 small models ready to run on end-user devices. It aligns with their Copilot+PC initiative, extending AI assistant experience to personal computers and tablets with Windows. 🔹 Azure Cobalt Chips: Azure Cobalt ARM-based processors are complamenting previously announced Maia custom AI chips, helping to reducing dependence on external hardware vendors. 3️⃣ 𝗚𝗼𝗼𝗴𝗹𝗲 𝗜/𝗢: 𝗦𝘁𝗿𝗲𝗮𝗺𝗹𝗶𝗻𝗲𝗱 𝗚𝗲𝗺𝗶𝗻𝗶 𝗺𝗼𝗱𝗲𝗹𝘀 𝗮𝗻𝗱 𝗧𝗿𝗶𝗹𝗹𝗶𝘂𝗺 𝗔𝗱𝘃𝗮𝗻𝗰𝗲𝗺𝗲𝗻𝘁 https://shorturl.at/J0ah6 🔹 Gemini Flash, a lighter version of Gemini 1.5 Ultra featuring a 1M token context window (soon up to 2M) offers tp developers better performance and lower costs. 🔹 Gemini Nano - emphasizing the potential of AI runing locally on mobile devices, Google introduced their smallest language model specifically designed for on-device AI applications. 🔹 6th generation of Trillium TPU promises a 5x performance improvement for AI workloads within data centers, and reduces reliance on Nvidia GPUs. We can expect more innovation when clients will be moving from testing and pilots to production.

5

At #T3DD24, I talked about the architecture of the mittwald cloud platform, the operational challenges that we encountered, and our future plans for ensuring long-term reliability and scalability. I've finally gotten around to posting a (two-part) write-up of said talk on the mittwald blog, brought up-to-date with everything that's changed in the past few months and a few more words on our plans for the very immediate future (in german; translation provided by the AI assistant of your choice). https://lnkd.in/gTpQBEwb

22

3 Kommentare

You can now deploy Oracle TimesTen In-Memory Database in #Kubernetes clusters on-premises or on #OCI Kubernetes Engine. Learn more: https://lnkd.in/gmixM24q

1

The Dynatrace Distributed Tracing app offers a new approach to handling OpenTelemetry data for teams. It integrates the extensive data collection capabilities of OTel with the Dynatrace platform, providing exceptional insight into the performance of your applications. In his blog, Florian Bacher explores the ways Dynatrace enables you to fully leverage your OpenTelemetry setup.

6

API Gateway + Kubernetes 🤝. Kong Inc. Champion and Chief Integration Evangelist / Chief Architect at OPITZ CONSULTING Deutschland GmbH, Sven Bernhardt on how to secure APIs with plugins in your Kubernetes management.  Full blog + video here covering:  - 🛠️ Installing Kong Ingress Controller - 🔐 Securing APIs with plugins - 📡 Exposing Kubernetes services as APIs, securing APIs with plugins (authentication, rate-limiting) - 🌐 Using the Gateway API for HTTP/TCP routing - 📊Observing analytics - all managed via the centralized Kong Konnect control plane. #KongPin #Kubernetes #Ingress #K8s #LearnToCode

5

Join my presentation tomorrow at 15:15h at #AWSSummit Berlin: "Leveraging Gen-AI to Enhance Augmenting Data-Driven Responses"

13

1 Kommentar

We wrote about the migration project from the on-premises environment to #AWS public cloud ☁️ solutions at Ringier Axel Springer Tech some time ago. Causes and challenges from the perspective of the entire organization – not just the technology stack 🌐💻, and on the other hand, the functionalities we were able to take advantage of. 🔀 This time we will dive 🤿 a little bit deeper and take a look at the data area. How have the organization’s needs and technology stack changed over time ⌛? What have we struggled with during the migration to the cloud and what are we currently working on 🧠? Read the full article „Data Warehouse - from in-house Solutions to the Public Cloud Reality” by Kamil Biesik and Krzysztof Golabek - our data engineers, on our blog: https://lnkd.in/deHBXBVT What are your experiences 💭 with migrating 🏃data 💽 to the cloud? #DataWarehouse #CloudMigration #RingierAxelSpringerTech

7

ICYMI, Google Cloud Platform and Technical Infrastructure Director, Jennifer Petoff, gave the opening keynote at DevOpsDays Zurich this year. You can check out the video here if you want to learn more about the importance of education in driving an SRE or DevOps culture.

3

#devops are table stakes for most dev teams and most have sophisticated in-house #devsecops tools and processes . With #googleclouds new #referencearchitecture, customers can now modernize software delivery operations further by integrating Google #genAI products into those tools and processes. Schneider Larbi and David Peterside recently published a great article - check it out #googlecloud #devops #genai #enterprisecomputing #distributedcomputing https://lnkd.in/gaVxF_Hb

28

Experience the full power of the Dynatrace platform on Microsoft Azure with Azure Native #Dynatrace Service, where customers can tap into the advanced AI capabilities of Dynatrace directly. Christoph Hoyer and Pawel Siwek highlight how our platform, combined with the #Azure Native Dynatrace Service (and features like Grail and Dashboards), provides instant insights into all the Azure resources deployed in your Azure tenant. Learn more here: https://dynatr.ac/4bqjNkG

46

Claude 3.5 Haiku is out on #AmazonBedrock following earlier announcements. It is a blazing fast model and surpasses even Claude 3 Opus in a number of benchmarks, but it starts out as a text only model. Anthropic had teased that the vision capability of this model is coming soon. Maybe as an update to the model. #aws #generativeai

16

🔍💡 Busting myths about platform engineering is crucial for understanding its true value and potential. In this article, our experts debunk five (more) common misconceptions, including the belief that platform engineering eliminates the need for infrastructure teams and that it slows down development. By clarifying these myths, the article highlights how platform engineering streamlines processes, enhances developer productivity, and supports robust infrastructure management. https://ow.ly/Jy6v50SLVAs #PlatformEngineering #DevOps #Infrastructure

4

AWS has just unveiled a game-changing solution for building cutting-edge generative AI applications: Real-time Vector Embedding Blueprints for Amazon MSK. 🔑 Key highlights:
 • Seamlessly integrate real-time data streams into your AI applications
 • Automatically generate vector embeddings using Amazon Bedrock
 • Index embeddings in Amazon OpenSearch Service
 • Simplify Retrieval Augmented Generation (RAG) architecture
 • Low-code approach to integrate MSK streams with OpenSearch and Bedrock This innovative blueprint addresses a critical challenge in generative AI: keeping models up-to-date with the latest information. By leveraging streaming data from Amazon MSK, you can now create AI applications that provide accurate, timely, and contextually relevant responses. The solution combines the power of:
 • Amazon MSK for real-time data streaming
 • Amazon Managed Service for Apache Flink for data processing
 • Amazon Bedrock for vector embeddings
 • Amazon OpenSearch Service as a vector database Whether you’re working on chatbots, content generation, or data analysis, this blueprint offers a streamlined path to building more responsive and accurate AI applications. Check out our blog post for a detailed walkthrough and start building your next-gen AI solutions today! #AWS #GenerativeAI #DataStreaming #MachineLearning #CloudComputing #AmazonMSK https://lnkd.in/gPWsCqW7

25

2 Kommentare