Andrew Hopper’s Post

The pace of GenAI model releases is still relentless. If you're not building AI systems today that enable you to change the model as the choices evolve, you either have an incredibly tight usage profile that is narrowly focused, or it's time to reconsider your architecture. While you're at it, ensure you include some automated testing to ensure your application functionality isn't regressive with each passing model change. What an incredible time it is to be building!

🔥 Releasing `llama-agents` and LlamaCloud (waitlist) 🚀 We’re thrilled to announce the release of `llama-agents`, our new open-source framework that simplifies the creation, iteration, and deployment of multi-agent AI systems, turning your agents into production-ready microservices. Key Features of `llama-agents`: 1️⃣ Distributed Architecture: Each agent operates as an independent microservice. 2️⃣ Standardized Communication: Seamless interaction through a central control plane. 3️⃣ Flexible Orchestration: Define explicit flows or rely on our smart orchestrator. 4️⃣ Easy Deployment: Effortlessly launch, scale, and monitor agents. 5️⃣ Scalable Performance: Monitor system and agent performance with our observability tools. Check it out here: https://lnkd.in/gD3EmcsW Additionally, we’re excited to introduce LlamaCloud, a fully managed ingestion service now open for waitlist: https://bit.ly/llamacloud 

Watch this on-demand webinar with Dynatrace's Bob Wambach and ESG's Jon Brown to learn how AI and Observability are becoming essential to transforming DevOps and IT operations through automation.

Been thinking a lot lately about the next evolution in AI and microservices: Domain-Oriented Microservice Agent Architecture (DOMAA). It’s about embedding AI agents directly into microservices to create systems that are intelligent, adaptive, and actually useful—not just hype. Especially in highly regulated industries, the key is balancing innovation with oversight, scalability, and trust. Excited to see how DOMAA can solve real-world challenges without sacrificing security or transparency. More to come! 🚀

New Kubernetes Release: AI Workload Efficiency and Other Game-Changers in Kubernetes v1.31: Elli Kubernetes 1.31 ushers in a new era of container orchestration, fortified by substantial advancements in AI/ML and networking capabilities. Kubernetes 1.31 exemplifies that the impact of a release — reflected in the contributions from 113 companies and 528 individuals — often surpasses the version number itself. This release demonstrates how point releases can meticulously refine the system for improved stability and #performance, while major releases continue to introduce groundbreaking capabilities that push the boundaries of what Kubernetes can achieve. #kubernetes #devops #cloudnative #k8s #Kubernetesv131Elli

Google has launched Gemini 1.5 - based on Mixture of Experts (MoE) architecture Gemini 1.5 is more efficient to train and serve, with a new Mixture-of-Experts (MoE) architecture. The first Gemini 1.5 model being released is Gemini 1.5 Pro. It’s a mid-size multimodal model, optimized for scaling across a wide-range of tasks, and performs at a similar level to 1.0 Ultra. It also introduces a breakthrough experimental feature in long-context understanding. Gemini 1.5 Pro comes with a standard 128,000 token context window. But starting today, a limited group of developers and enterprise customers can try it with a context window of up to 1 million tokens via AI Studio and Vertex AI in private preview. Details here -> https://lnkd.in/dab6GCSt _____________________________________ Follow me for regular updates and insights on Generative AI and my journey as a CTO in AI product development.

Ever struggled with the complexity of managing large-scale applications? That's where microservices architecture comes into play. By breaking applications into smaller, independent services, microservices allow for easier management and scalability. But the game gets even more interesting with the integration of AI. AI can dynamically allocate resources, predict failures, and streamline communication between services, making the system more robust and efficient. In my personal experience, implementing AI with microservices has drastically reduced system downtime and improved overall performance. Are you exploring microservices? How has AI impacted your architecture? #webevolution #webdeveloper #ai

Unlike regular platform engineering, which focuses primarily on foundational IT and infrastructure, Platform Engineering++ integrates DevOps and infrastructure roles with support for advanced areas like data, machine learning, AI, orchestration, and APIs. Learn more about this new approach to building the modern data platform by Luca Cruciani of Mia-Platform. https://lnkd.in/dbpY6Ci5

Agile and Software Architecture have always had a complicated relationship 🤯 AI is making it even more complicated... or not? 💡 Is the role of the architect destined to evolve 🧬 ...to be an AAA? (Agile and AI Architect). I talked about this and our study conducted with Jan Bosch on CAFFEA and agile architects at #ITARC 2024 in Stockholm. Thanks Thijmen de Gooijer and Rebekka Wohlrab for inviting me to talk at #ITARC, great event organized by DF Kompetens with many architects and interesting presentations and workshops. #agile #softwarearchitecture #agilearchitecture #softwarearchitect #AI

After studying how companies deploy generative AI applications, I noticed many similarities in their platforms. This post outlines these common components, what they do, and implementation considerations. This post starts from the simplest architecture and progressively add more components. 1. Enhance context input into a model by giving the model access to external data sources and tools for information gathering. 2. Put in guardrails to protect your system and your users. 3. Add model router and gateway to support complex pipelines and add more security. 4. Optimize for latency and costs with cache. 5. Add complex logic and write actions to maximize your system’s capabilities. I try my best to keep the architecture general, but certain applications might deviate. As always, feedback is appreciated!