Sharon Resheff

This is one you should not miss! Tony Seale ‘s insights with Morgan Templar will change the way you think about organising, sharing and describing data to improve your enterprise.

21

New research from the Databricks AI research team shows that automated prompt optimization can deliver quality improvements on par with supervised fine-tuning—and with lower serving costs.

5

I wrote this evaluation of the role of analytics engineering in the modern data ecosystem to help clarify the value prop of a well differentiated team. Enjoy!

14

If you're evaluating semantic layers and Strategy Mosaic isn't on your list, this is worth a watch. Igor Freitas put together a side-by-side comparison with dbt, Cube, and AtScale, covering governance maturity, AI capabilities, data source breadth, and BI connectivity. 👉 https://lnkd.in/dh5YnmfJ

59

6 Comments

Remember when demand forecasting required complex SQL and manual model training? Those days are ending. BigQuery is moving from a "Builder" approach to an "Agentic" approach using Conversational Analytics and the TimesFM (Time Series Foundation Model). Check out this article from Karl Weinmeister to learn more: https://lnkd.in/gCFWu78Y #conversationalanalytics #agenticAI #MCP #GoogleADK #VertexAI #timeseriesforecasting

11

🔥 Snowflake Feature Spotlight: Streams + Tasks = Real-Time Data Pipelines One of the most powerful (and underrated) capabilities in Snowflake is how easily you can build near real-time pipelines — without external orchestration tools. Here’s a simple pattern I’ve been exploring: 👉 Streams → Capture CDC (INSERT / UPDATE / DELETE) 👉 Tasks → Automate processing (scheduled or triggered) 👉 Merge Logic → Maintain SCD Type 2 history Why this is powerful: ✔ No need for external tools like Airflow for many use cases ✔ Built-in change tracking (no complex CDC setup) ✔ Scalable and serverless execution ✔ Perfect for incremental data processing Real-world use case: Imagine ingesting customer updates from multiple systems — using Streams + Tasks, you can: → Capture every change → Process only new data → Maintain full historical tracking (SCD2) All inside Snowflake. This is where Snowflake really shines — simplicity + power. 💡 Curious: Are you using Streams & Tasks in production, or still relying on external schedulers? #Snowflake #DataEngineering #Streams #Tasks #SCD2 #ModernDataStack #ELT

6

Moving from AI Experimentation to Execution has it's challenges. Listen to this podcast to glean what you need to know to make it happen at your organization.

1

We are witnessing a transformative shift in how decisions are made and value is created, driven by the acceleration of AI. 🌐 As data becomes more accessible, the demand for trusted insights has never been greater. I’m thrilled to share two exciting announcements from S&P Global that will empower our customers on their AI journey: 1️⃣ Integration with Anthropic We have officially brought S&P Global's trusted financial data into Anthropic's Claude! 🤝 This collaboration enhances the accessibility of our insights, allowing financial professionals—from hedge fund managers to private equity analysts—to obtain reliable answers to complex questions in real-time within the GenAI ecosystem: https://lnkd.in/eCgujYvx 2️⃣ Launch of AI-Ready Metadata: Our new AI-ready metadata on the S&P Global Marketplace marks a significant leap in how our customers consume data in an AI-first world. 📊 This provides machine-readable metadata that accelerates integration into AI workflows. By delivering structured, interoperable, and context-rich data, we are not only meeting the rising demand for actionable insights but also ensuring our data is ready for AI from the start: https://lnkd.in/eC5By2ms In an era where AI adoption is rapidly transforming workflows, the importance of trusted insights cannot be overstated. 🔍 At S&P Global, we are committed to delivering intelligence that empowers our customers to make informed decisions with clarity and confidence!

745

9 Comments

dbt Labs + Tableau 🤝 Making governed dbt metrics available directly in dashboards, reducing drift, and improving performance. Read more: https://lnkd.in/g73J2Nvf

4

Excited that Cortex Code CLI is now generally available. Cortex Code is the Snowflake-native AI coding agent that turns complex data workflows into simple interactions—all in natural language. Whether you’re in Snowsight or working via the CLI, Cortex Code is deeply aware of your Snowflake environment, meaning Snowflake expertise becomes an always-on capability of the platform—for everyone. Dive in to learn more: https://lnkd.in/dxEQpFYm

2

Integrated vector databases are now the preferred solution for enterprise AI retrieval. Teradata’s extension of Enterprise Vector Store adds multimodal and agentic AI capabilities. It reflects a broader shift IDC has been seeing over the past year, with organizations increasingly favoring vector capabilities built into existing data platforms over standalone native vector databases. Why? Simpler architectures, less data movement, and more consistent governance as agentic AI moves into production. IDC Link https://lnkd.in/e4HJbY-C #IDC #Teradata #VectorSearch #AgenticAI #EnterpriseAI

71

11 Comments

Last week we announced the launch of @Snowflake native semantic views at #snowflakesummit.  Snowflake Semantic Views make it easy to define consistent metrics and dimensional context—right inside Snowflake. Whether you're working with BI tools, AI agents, or just trying to reduce logic drift across teams, Semantic Views provide a governed, SQL-native layer that scales. Feedback from customers and partners has been very positive, and there are lots of questions on how to get started. To help you get hands-on, we’ve launched a Semantic Views Quickstart. It walks you through: 🛠️ Modeling — Defining semantic models using SQL 📊 Metrics — Creating measures with built-in dimensional awareness 🔍 Querying — Exploring semantic views directly or with Streamlit 💬 Natural Language — Asking questions with Cortex Analyst Link here: https://lnkd.in/gMYGWRHf If you're evaluating how to centralize and scale your semantic layer within your existing Snowflake environment, this is a great place to start. Try it out and let us know what you build. #Snowflake #SemanticLayer #DataModeling #BI #Analytics

116

No AI Strategy Without a Data Strategy I spent yesterday at the Snowflake World Tour in Chicago, and it reinforced what I’ve been saying from the beginning, a strong AI strategy is only possible with a strong data strategy behind it. Too often, organizations jump straight to “AI strategy” (models, copilots, agents) without shoring up the foundation. But a good AI strategy starts with a good data strategy. Here’s what that looks like in practice: Unified Data Infrastructure: Eliminate silos. When data is scattered across systems and formats, even the best models will hallucinate, drift, or fail. Data Quality & Governance: Garbage in, garbage out. Consistent cleaning, labeling, and governance ensure that insights (and AI outputs) can be trusted. Accessibility with Control: The right people, and the right models, need frictionless access to accurate, secure data. Measurement & Feedback Loops: Observability isn’t just for models. You need continuous visibility into how data is collected, transformed, and used. AI isn’t magic. It’s math on data. If your data house isn’t in order, your AI roadmap will crumble no matter how sophisticated the architecture looks. Is your data strategy strong enough to support the weight of your AI ambitions? #AI #DataStrategy #Snowflake #DigitalTransformation #EmergingTech #Analytics #Leadership

272

56 Comments

I think Databricks’ Genie Code is clearly less about introducing an entirely new product and more about formalizing the steady improvement of its AI development experience over the last year. Not only has the output quality improved over time, but features such as dashboard authoring, Lakeflow Spark Declarative Pipelines support, etc were under-the-radar additions that positively impacted development speeds, quality, and time-to-delivery. Some competitors have had to rethink or replace rough first-generation coding products, Databricks started with something usable and kept improving it, so hey, might as well celebrate with a new name too! Also, I believe the emphasis on more agentic capabilities points to where the platform is going next, especially as Databricks pushes toward a more unified AI product experience built around governed enterprise data.

200

11 Comments

🎯 Andy Konwinski kicked off Databricks #DAIS Dev Night spotlighting the need for benchmarks supported by hard human data. We couldn't agree more. Agentic AI and GenAI applications will not be deployed for mission critical use cases unless they become as trusted as human experts. There's a massive shift underway from model inputs to AI evaluation, outcomes, & consequences. In this new world, out-of-the-box generic benchmarks and lightly customized evals don't cut it. At Snorkel AI, our mission is to help enterprises transform their valuable expert knowledge and IP into specialized AI at scale. You've got questions? We've got answers. ⏩ Join Chris Borg Wed at 2pm to learn how Snorkel helps F500 Databricks users "Evaluate and Improve the Performance of Agentic Systems" ⏩ Join former Wells Fargo exec Maxwell Williams and client success lead Gabe Smith for Breakfast with Data Scientists Thursday ⏩ Visit Snorkel's booth D116 #agenticevaluation #DataAISummit #agenticai

19

1 Comment

#semantic_layer is key to building the data architecture layer for the enterprise. If you think you can build business insights just using an LLM call you should explore here first.

6