Jules Damji

Very interesting post on Intuit's use of LLMs to manage K8s, via Aditya Kulkarni at InfoQ: "Intuit Engineering's Approach to Simplifying Kubernetes Management with AI" https://lnkd.in/eRN6_sAK

2

Learn the ins and outs of data latency with this expert guide from Oracle. https://lnkd.in/gnQJdJgS

1

In 2013, Jay Kreps introduced the concept of the log as a fundamental abstraction for real-time data streaming, paving the way for platforms like #ApacheKafka. Over the past decade, the data streaming landscape has undergone a remarkable transformation, particularly with the rise of cloud-native environments and the integration of scalable object storage solutions like Amazon S3. This evolution has given rise to platforms like #ApachePulsar, each offering innovative approaches to log storage and data streaming. In our latest blog post, we explore the evolution of log storage implementations in data streaming and categorize this architectural journey into three distinct stages: 💡 Leader-Based Architecture Leader-based architectures form the foundation of most data streaming technologies used today. Traditional message queues rely on single-node brokers with primary-backup setups for high availability (HA). Kafka was the first distributed implementation, but it remains leader-based, where a designated leader manages replication. Pulsar advanced this by decoupling compute and storage, but it still follows the leader-based model. This approach works exceptionally well for scenarios requiring ultra-low latency (e.g., milliseconds). For such use cases, sending data through a leader remains unavoidable. 💡 Leaderless Architecture The emergence of S3 as a primary storage layer has triggered a shift toward leaderless architectures. Instead of treating S3 as just a tiered storage solution, it now serves as the main storage layer. In this model, brokers become leaderless, meaning any broker can handle reads and writes. Platforms like WarpStream, Confluent Freight Cluster, and StreamNative’s Ursa Engine are adopting this approach, reflecting the growing trend toward distributed and flexible architectures. 💡 Brokerless / Headless Architecture The next frontier is headless (or brokerless) architectures, where metadata and data are entirely decoupled: Data resides in S3, while metadata is managed by a separate metadata or catalog service. With a standard specification in place, operations no longer need to flow through brokers. Vendors can simply adhere to the standard and build their solutions accordingly. This approach can enable seamless scalability and interoperability. Our Ursa Engine is heading toward this exciting future. Explore the full blog post to dive deeper into these trends and the journey from leader-based to brokerless architectures: https://lnkd.in/gCcmv3rU You can also watch our keynote presentation at Data Streaming Summit: https://lnkd.in/g6a3wHrC

61

Fine tuning an LLM (Large Language Model) means adapting it to specific data and tasks with the goal of improving performance on those and similar tasks. In a new blog post, my colleagues describe a set of best practices for fine-tuning Anthropic Claude 3 Haiku on Amazon Bedrock: 1 - Choose a suitable use case per the guidelines in the post. 2 - Gather a clean, high-quality data set and structure it for efficient training, with an optional system prompt. 3 - Customize and run the fine-tuning job. 4 - Assess the fine-tuned model's performance against relevant metrics and against the base model. 5 - Iterate and deploy. Read the entire post at https://lnkd.in/gAT6kzHK to learn a lot more!

121

9 Comments

Since we launched Redshift Streaming Ingestion with Kinesis Data Streams (KDS) and Managed Streaming for Apache Kafka (MSK) we have seen great adoption with thousands of customers ingesting data at very high rates with low latency directly to Redshift Managed Storage. One common request was to extend the ingestion capabilities to other Kafka sources. Today we are extending Redshift's Streaming Ingestion to support self-managed Apache Kafka on EC2 as well as partners, such as Confluent Cloud. Go and Ingest! https://lnkd.in/gpzC-JD2 #amazon #redshift #streaming #ingestion #kafka #confluent

200

1 Comment

Here is a blog post that explains how to use OCI Generative AI to perform Retrieval Augmented Generation. The article focuses on various hyper-parameters that can be tuned. https://lnkd.in/ghP5r6ef #GenerativeAI

34

1 Comment

Oracle today announced the general availability of HeatWave #GenAI, including in-database LLMs, an automated in-database vector store, and the ability to have human-like contextual conversations in natural language. https://lnkd.in/gwZc7Pi3

3

Redis is fueling the data revolution with Google Cloud Marketplace! ⚡ In a recent episode of The Ultimate Guide to Partnering, Ash Vijayakanthan, SVP Cloud Sales and BD Partnerships, Redis shared how they're leveraging the cloud marketplace to accelerate procurement and scale high-performance data solutions for customers. With the increasing demand for real-time data and the rise of #GenAI, high-performance caching and database solutions are more critical than ever. Their marketplace strategy is not only helping them retain customers but also expand deal sizes and accelerate closing times. 🚀 Key takeaway: Investing in the marketplace and empowering your sales teams can unlock significant business value. Don't miss out on this insightful conversation! #GoogleCloudPartners Ultimate Partner

4

Watch the final part of the series "Crafting Applications with Cosmos DB" where Saranya Sriram and Sandeep S Nair unpack how surprisingly simple it is to make your app on Cosmos DB resilient in the case of a Datacenter failure. They also refer to a new preview feature that is a steal: Dynamic Autoscaling: https://lnkd.in/e74PXE4V #Autoscale Azure Cosmos DB #HighAvailability #NoSQL #DistributedData https://lnkd.in/e8he9PCy

18

As we enter the #AI era #systemsarchitecture has to adapt to the new reality of #hybridcloud and #multicloud. Wyatt Gorman HPC Solutions Manager for Google Cloud joins me in the first episode of #partlycloudy to discuss #manageability challenges related to #aiinfrastructure https://lnkd.in/ga25iFFk

6

1 Comment

I'm excited to share our latest blog on deploying an LLM inference endpoint on AWS SageMaker with Friendli Container. If you're already using SageMaker, give it a try. #friendliai #friendli #container #aws #sagemaker #llm #inference #endpoint

13

This Oracle #HeatWave #GenAI demo will show you how you can have contextual conversations in natural language informed by your unstructured content. https://lnkd.in/gJbAXEND

3

Developers love Redis because it’s fast. Liftoff uses Redis Cloud as a high-availability, low-latency database, increasing their ROI by improving ad relevance and quality. Hear more about the benefits of their improved ad relevance and quality straight from the director of engineering at Liftoff Mobile. Learn more about Liftoff's soaring ad exchange with Redis Cloud's low latency: https://lnkd.in/gAWBScEk

36

1 Comment

One of the most powerful capabilities that Redshift is launching today, Amazon Redshift now supports incremental refresh for Materialized Views (MVs) on data lake tables. With IVM on Data Lake MVs users can now combine the strengths of data lake and high performance of warehouses.

187

8 Comments

A really nice read from Sai Maddali on the benefits of Tiered Storage with Amazon MSK. https://lnkd.in/gGNJ5nyt #ApacheKafka #Kafka #AWS #MSK #tieredstorage #streaming #storage

55

2 Comments