Evangelos Pertsinis

Last day of the Databricks Data+AI summit and Nexla is showing off its Advanced #RAG framework for #GenAI. Come get a demo! - Complete out of the box framework, ready to use as a chatbot or API endpoint - Works with Hybrid data: Vector Databases, API feeds, Databases - Complete with Query rewriting, Re-ranking, Monitoring. See how you can take your RAG prototypes to production with Nexla

31

1 Comment

Exciting news! Snowflake has unveiled a cutting-edge system for large language models, featuring the powerful Llama 3.1 405B from Meta. This innovation brings a new level of efficiency and capability to AI applications. At 7Rivers, we're excited to harness this technology, enhancing our Data-Native™ approach to deliver unparalleled AI solutions for our clients. Our commitment at 7Rivers is to drive business value through use of intelligent data and advanced AI. Snowflake's latest offering, which allows real-time AI processing with minimal infrastructure, perfectly complements our mission. This advancement enables us to provide our clients with scalable, cost-effective AI applications, enhancing operational efficiency and driving impactful outcomes. The integration of Meta’s Llama 3.1 405B within Snowflake Cortex AI aligns with our Data Native™ Model, empowering us to create transformative solutions, optimizing processes, and creating a competitive advantage. Together with Snowflake, we're revolutionizing data usage to create intelligent, augmented business experiences. #AI #DataNative #7Rivers #Snowflake #GenerativeAI #BusinessTransformation

33

2 Comments

🎉 29 Days of Airflow 2.9: Day 23 🎉 📈 Today's feature update in #ApacheAirflow 2.9 focuses on enhancing data visibility: adding datasets to the DAG graph. This improvement, delivered via pull request #37604, integrates datasets directly into the visual #DAG graph in the #Airflow UI. This allows users to see how datasets are connected to various tasks, providing a clearer understanding of how data flows through and impacts workflows. 🔍 Why is this significant? Incorporating datasets into the DAG graph helps users visualize and comprehend the dependencies and relationships between tasks and data more intuitively. This addition is crucial for troubleshooting and optimizing data pipelines, ensuring a more efficient and effective data handling process. 🔗 Learn more about the benefits of visualizing datasets in your DAG graphs: https://bit.ly/458xvqh #Airflow29 #DataVisualization #29daysofairflow29 #DataEngineering

19

One of the big annoucements at the Databricks Data + AI Summit was their AI/BI Genie. Took some time yesterday to do a side-by-side comparison on whats involved to get it working, and how well it performs vs. Zing Data. We conducted a three-step test to evaluate both platforms: 🚚 Data Loading - Zing Data: Upload a single Excel file with multiple sheets - Databricks: Upload individual CSV files for each Excel sheet tab ✍ Data Preparation - Zing Data: Immediate querying capability - Databricks: Manual setup of Unity Catalog source and Genie Workspace, with individual table specification 📊 Data Querying and Visualization - Zing Data: Direct graph manipulation, visual addition of splits, colors, and filters - Databricks: Changes require SQL modifications 👉 Results - Zing Data completed the tasks in 1 minute 44 seconds - Databricks completed the tasks in 7 minutes 39 seconds This test showed Zing Data performing 4.4x faster than Databricks for this end-to-end loading + querying + dataviz task. 👀 Key Observations - Zing Data offers a more streamlined data loading process, allowing multi table imports from a single file and not needing to specify specific tables to query ahead of time. - Zing's interface allows for direct visual data manipulation to perform sorts, filters, and additional splits if the AI results were not what you wanted. Joins could be added / modified visually. - Databricks requires more manual setup and SQL knowledge for similar tasks (e.g. applying a sort or filter) 💲 Pricing and Getting Started Zing Data offers a free tier, allowing users to test the platform. Databricks' AI/BI offering is not available through a free trial or community edition. It requires a Premium tier / Serverless SQL subscription to kick the tires. 💻 For those interested in exploring Zing Data, visit www.zingdata.com #DataAnalysis #BusinessIntelligence #PerformanceComparison #databricks #zingdata #genai #datavisualization #dataanalytics #datascience

16