Sheng-Fu Chuang

See how simple it is to create chatbots with Microsoft Azure On Your Data and vector embeddings from an existing Elasticsearch index: https://lnkd.in/gBHZbkjq #ElasticSearchLabs #RAG #VectorDatabase

2

🚀 Building Scalable ML Inference Pipelines Great article by @Simardeep Singh showing how to combine: - Dapr for resilient microservices - Knative for serverless scalability - Ollama & Llama 2 for ML inference Perfect for teams looking to build production-ready ML pipelines with cloud-native best practices! See how these powerful tools work together: https://diagrid.ws/3YCIk23 #MLOps #CloudNative #Dapr #Knative #MachineLearning

1

Looking to save time and gain agility? Discover how embracing #klaw helped Essendant do just that in our latest case study 👉 https://ow.ly/vfXt50Q1ZxQ #klaw #apachekafka #casestudy

6

Implement a Q&A experience using a RAG technique with Elasticsearch as a vector database by following the steps in this blog: https://gag.gl/V2hwuA #ElasticSearchLabs #RAG #GenAI

2

Building Event-Driven Architectures with Apache Kafka In today’s fast-paced digital landscape, real-time data processing is crucial for modern applications. At TrueulyMarket, we leverage Apache Kafka to build robust and scalable event-driven architectures that enhance the performance and reliability of our applications. Apache Kafka is a powerful stream-processing platform that enables us to handle real-time data feeds efficiently. Here’s how we utilize Kafka to create event-driven architectures: 1. Real-Time Data Processing: Kafka allows us to process and react to data in real-time, ensuring that our applications can handle live updates and events seamlessly. 2. Scalability: With Kafka’s distributed architecture, we can scale our applications horizontally to handle increased data volumes and user loads. 3. Reliability: Kafka’s fault-tolerant design ensures high availability and reliability, making it an ideal choice for mission-critical applications. 4. Decoupled Services: By using Kafka, we can decouple services and components, enabling independent development, deployment, and scaling of each service. 5. High Throughput: Kafka’s ability to handle high throughput makes it perfect for applications that require processing large streams of data quickly and efficiently. 6. Durability: Kafka ensures data durability through its distributed log storage, guaranteeing that no data is lost even in the event of system failures. 7. Integration: Kafka integrates seamlessly with various data sources and sinks, allowing us to build flexible and interconnected systems. 8. Enhanced Analytics: With Kafka, we can build powerful real-time analytics and monitoring solutions, providing actionable insights from live data streams. At TrueulyMarket, our expertise in Apache Kafka and event-driven architectures enables us to build high-performing, scalable, and reliable applications. Partner with TrueulyMarket to leverage the power of real-time data processing and enhance your digital solutions. #TrueulyMarket #ApacheKafka #EventDrivenArchitecture #RealTimeData #ScalableSolutions #TechInnovation #DigitalSolutions #DataProcessing #HighThroughput #DecoupledServices #Reliability #FaultTolerance #TechCommunity #SoftwareEngineering #ModernDevelopment #TechExcellence #DevelopmentTrends #TechSavvy #DigitalInnovation #AppOptimization #UserExperience #DataIntegration #Analytics #StreamProcessing #FutureOfTech #KafkaDevelopment #AppEnhancements #ScalableArchitecture #RealTimeAnalytics #Durability #EventDrivenDevelopment

1

Unlock Efficient Resource Management with #Terraform and #meshStack! 🚀 Utilize our Terraform provider to manage projects, user assignments, and tenants securely and effectively. Fully leveraging infrastructure-as-code is key in the cloud-native era. Application Teams using meshStack can use our Terraform provider to fully manage their meshStack resources using Terraform. 👉 Manage your projects ✔ 👉 Manage user assignments ✔ 👉 Manage your tenants ✔ & more support for the Terraform provider is to come. 🧠 The example above shows how you can create (or update) a project in meshStack using Infrastructure-as-Code. This allows application teams to manage their resources in meshStack more securely and efficiently as opposed to managing these via the user interface. #InfrastructureasCode

9

Curious about how we manage over 300 Kafka topics across 9 clusters with ease? 🤔 Head over to the LSports Engineering Blog to read Nizan Rosh's latest article. Learn about "Kafka-Curator," our open-source tool that simplifies Kafka topic management while boosting performance and reliability. Discover the secrets behind our scalable and resilient infrastructure. Don't miss out! read more at LSports Engineering Blog >> https://lnkd.in/eJUPpZcM

34

2 Comments

Managing Kafka + Debezium for real-time data streaming can be an operational burden, especially at an enterprise scale. It involves complex configurations, ongoing maintenance, and ensuring security and reliability across multiple systems. For organizations that need a more streamlined and secure solution, Estuary Flow offers a robust alternative. With Estuary Flow, you get enterprise-ready real-time data integration without the hassle. It can be deployed in your own private environment, ensuring full control over your data while benefiting from the speed and efficiency of real-time connectors. Simplify your architecture and reduce the operational overhead with a secure, scalable solution built for modern data needs. Reach out directly if you are interested: https://lnkd.in/gwnnmVNR Join the conversation on Slack: https://lnkd.in/eQ_hRGKW

24

My research into cell-based architecture led me to look into the latest trends in #SoftwareArchitecture So here they are (some of these are already quite common); 1. Data-driven architecture 2. Edge computing 3. Platform architecture 4. Large language models 5. Socio-technical architecture 6. Green software 7. Privacy engineering Fascinating reading up about these and will definitely be weaving them into conversations to add some meat to the bones. #everydayisaschoolday

16

3 Comments

#Profesia per la #digitaltransformation:WSO2’s API Platform for Kubernetes Open Source Future with Sanjeewa Malalgoda Join WSO2’s Director of Engineering, Sanjeewa Malagoda, as he discusses WSO2’s commitment to open source software and our newest open source release: API Platform for Kubnetes (APK). Topics will include: - The future of API management. - Why WSO2 built the API Platform for Kubernetes...

1

Learn about the evolution of the Elasticsearch .NET client and how it will gradually phase out the old NEST library: https://gag.gl/tCCrQX #ElasticSearchLabs

1

CEO Spotlight: Lukas Gentele, CEO, Loft Labs: Creators of vCluster. $24million Series-A, April, 2024. ** Multitenancy in Kubernetes with vClusters ** * Why VCs are So Excited About vCluster? (13:03) (Remaining Time) * How vCluster Different from All Cloud Mgmt Tools? (11:56) * Why are vClusters More Secure? (10:06) * Why do vClusters = K8s Uptime and Resiliency? (9:24) * How big is the Future Demand for vClusters? (8:42) * How does vCluster make K8s Easier for Developers? (6:37) * New Partnerships? Rancher, Coreweave, Vault, TF, etc. (4:59) * What is the Ideal Customer Profile for vCluster? (3:40) * Preferred Method for People to Learn More? Slack. (1:04) #kubernetes #vcluster #loftlabs

37

9 Comments

Heading to #reinvent2024 next month? Don’t miss our hands-on workshop on Dec. 4th on scaling generative AI workloads with Anyscale & Amazon Web Services (AWS) Hands-on session includes: 🔹 Building end-to-end LLM applications with RayTurbo, Anyscale's optimized Ray runtime. 🔹 Applying advanced performance techniques like dynamic autoscaling, heterogeneous hardware integration (CPU/GPU), and replica compaction. 🔹 Fine-tuning large language models with Anyscale’s LLM-Forge solution. 🔹 Using Anyscale observability tools to debug, diagnose, and profile distributed AI workloads effectively. ....and more! Register now: https://lnkd.in/gqRYFPz6 Seats are limited—sign up today!

8

How LinkedIn Reduced Latency with JSON. This case study is a goldmine for anyone working on large-scale systems. The Challenge: -- 900M+ members across 200 countries -- 1000s of backend services, 10,000s of API endpoints -- JSON serialization causing high bandwidth usage and latency Initial Attempt: Tried compression algorithms like gzip. Result: Reduced payload size but worsened serialization/deserialization latency The Solution: After evaluating multiple alternatives, LinkedIn chose Protocol Buffers (Protobuf). Here's why: Protobuf Benefits: • Smaller payload (e.g., 2 bytes vs 10 bytes for simple messages) • Faster serialization/deserialization • Strong typing with predefined schemas • Wide language support • Backward compatibility for schema updates Alternatives Considered: • Flatbuffers: Offers "zero-copy deserialization" • MessagePack: No predefined schema is required • CBOR: Inspired by MessagePack, with additional features Results: 1. Up to 60% latency improvement for large payloads 2. Increased throughput for both request and response payloads 3. No significant degradations compared to JSON LinkedIn's Rest.li Framework: They've open-sourced Rest. li, a Java framework for RESTful services that handles: 1. Service Discovery (e.g., translating d2:// to https://lnkd.in/g9euZv8g) 2. Type Safety 3. Load Balancing 4. Common Request Patterns (e.g., Scatter-Gather) Key Takeaways: 1. Data serialization format significantly impacts system performance at scale 2. Binary formats like Protobuf can offer substantial benefits over text-based formats like JSON 3. Consider trade-offs between human readability, flexibility, and performance 4. Custom frameworks like Rest.li can simplify microservices architecture management This case study shows how seemingly small optimizations in system design can lead to massive performance gains at scale. What's your experience with data serialization in large-scale systems? Have you used Protobuf or similar binary formats? Let me know in the comments! #SystemDesign #TechOptimization #Protobuf #REST

817

28 Comments

How Terraform Powers Scalable and Reproducible Machine Learning on AWS Terraform has become an invaluable tool in managing and automating the cloud resources necessary for machine learning (ML) pipelines. Here’s how I leverage Terraform in AWS to streamline my ML processes: 1. Automating AWS Resource Provisioning: Machine learning workloads often require complex infrastructure, including GPU-enabled EC2 instances for training, S3 buckets for storing large datasets, and Elastic Kubernetes Service (EKS) clusters for distributed training or serving models at scale. With Terraform, I can define all these resources in declarative configuration files. This allows me to automate the provisioning of EC2 instances with specific instance types (e.g., p3.2xlarge), configure S3 for scalable object storage, and even set up Amazon SageMaker for model training and deployment—all with a single command. This automation simplifies scaling and reduces manual errors in resource setup. 2. Modular and Reproducible Infrastructure: In ML, reproducibility is key. Terraform allows me to create reusable modules for different parts of my infrastructure—like setting up VPCs, IAM roles, or EKS clusters. By breaking down my infrastructure into reusable modules, I ensure consistency across development, testing, and production environments. This approach makes it easy to recreate the same infrastructure anywhere in the world with just a few tweaks, which is crucial for running experiments or replicating models on different datasets or regions. 3. Integrating with CI/CD Pipelines: Terraform integrates seamlessly with CI/CD tools like AWS CodePipeline or GitHub Actions, allowing me to automate the deployment of my ML models and infrastructure. For example, when a new model is pushed to my GitHub repo, Terraform can automatically update the infrastructure—such as scaling up EC2 instances for retraining or updating the Lambda functions that serve inference requests through API Gateway. This ensures a smooth and repeatable deployment process, minimizing downtime and human intervention. Conclusion: Terraform, when paired with AWS, provides a robust and scalable way to manage machine learning infrastructure. From automating resource provisioning and cost management to ensuring security and reproducibility, Terraform enables me to focus on developing and deploying machine learning models with confidence. The ability to treat infrastructure as code not only streamlines the workflow but also reduces the operational overhead associated with managing cloud resources, allowing for quicker iteration and better results. #MachineLearning #Terraform #AWS #MLOps #DataScience #Automation #InfrastructureAsCode

11

1 Comment

Want to accelerate your software delivery while keeping costs in check? Read our blog to learn how to build a robust #AWS CI/CD pipeline that prioritizes both speed and security. In partnership with AWS Partners, we've developed a comprehensive blog covering: ✓ Automated testing with AWS CodeBuild and CodePipeline ✓ Containerization using Amazon ECS and EKS ✓ Infrastructure as Code with AWS CloudFormation ✓ Security best practices with AWS IAM and KMS Learn how to accelerate your software delivery while maintaining security and compliance. Read the blog now: https://hubs.li/Q02W2B2p0. #AWSPartners #APNproud #CICD

4

Explore potential bottlenecks and ways to mitigate growing pains around Elasticsearch inference pipelines when dealing with complex workloads, from downloading the model to monitoring it. Learn more here: https://gag.gl/AfNudQ #ElasticSearchLabs #ML

2

"Choosing the Right Local Kubernetes Development Tool: A Comparative Analysis of Telepresence, Gefyra, and mirrord" by Vocal Creators https://lnkd.in/gGrreFiY #mirrord #kubernetes

8