Timothy Spann

Posted on Apr 6

All Data and AI Weekly #236-06-April-2026

#ai #agentaichallenge #llm #mcp

All Data and AI Weekly #236-06April2026

( AI, Data, Agentic AI, Cortex Code, NiFi, Iceberg, Polaris, Streamlit, Python, Java, SQL, MCP, LLM, RAG, Cortex AI, AISQL, Search, Unstructured Data )

Tim's Corner: An Easter Basket Full of Data Goodness

Happy Easter and Happy Passover! Welcome to issue 236! Just like an Easter egg hunt, this week is packed with hidden treasures waiting to be discovered across the Snowflake ecosystem. The engineering team cracked open a must-read on why you need a semantic model, the Agentic Enterprise Control Plane vision continues to hatch new ideas, and Ontology on Snowflake went from architecture to deployment with a dedicated Cortex Code skill. On the platform side, CTAS support for Unity Catalog external volumes landed, Snowflake Directory Imports are here, and the Network Policy Advisor gives you a new egg-shaped lens on your security posture. Meanwhile, the community has been busy - guides on migrating Tableau to Streamlit with AI, building production-ready dbt projects with Cortex Code, and deploying Gemma 4 multimodal on SPCS. I've been building new SNACK-AI projects including an OSINT toolkit and SnowGhostBreakers Advanced. Spring is here - let's dig into the basket!

Tim's Content

Happy Passover and Easter.

Recent Presentations

Beyond the Prompt: Future of AI Data Cloud - March 2026 - Slide deck
24 March 2026 Demo Notes - NYC Snowflake Users Group hands-on demo

New SNACK-AI Repositories

Repository	Description
SNACKAI-CoCo-UnstructuredDemos	End-to-end demos for processing unstructured data with Cortex Code and SNACK-AI, including a dev ports example
SNACKAI-CoCo-BeyondthePrompt	Companion repository for the Beyond the Prompt presentation on the future of AI in the Data Cloud
SNACKAI-CoCo-Snowghost2	Ghost-busting application v2 built with Cortex Code on Snowflake
SNACKAI-CoCo-SnowGhostBreakers-Advanced	Advanced ghost-busting analytics with enhanced detection and investigation workflows
SNACKAI-Cortex-Code-Cheatsheet	Quick-reference cheatsheet for Cortex Code commands, skills, and workflows
SNACKAI-CoCo-OSINT	Open-source intelligence toolkit built with Cortex Code and SNACK-AI

Upcoming Events: Catch Tim Live!

April 7: Snowflake Connect: AI Data Cloud - Virtual
April 9: The Agentic Enterprise: From Strategy to ROI - Virtual
April 13-17: D4BV - Virtual APAC
April 14: D4BNA - Virtual US
April 15: Nodes AI Conference - Virtual
April 22: Snowflake Connect: Building Transformation Pipelines for AI-Ready Data - Virtual
April 29: Tysons Corner Hands-On Lab: Automate Data Workflows with Cortex Code and Evolv - Tysons Corner, VA
May 20: Budapest Data and AI - Budapest, Hungary - Tim Virtual speaking
June 2-5: Snowflake Summit 2026 - San Francisco, CA - Join me, I will be speaking on NiFi and Agents.

All Snowflake Events: https://www.snowflake.com/about/events/

Snowflake Summit 2026 - San Francisco, June 2-5

The biggest event of the year is coming! Snowflake Summit 2026 returns to San Francisco June 2-5. Expect keynotes, hands-on labs, deep technical sessions, partner showcases, and announcements that will shape the future of the data and AI ecosystem. Whether you're a data engineer, data scientist, architect, or business leader - this is the event to be at.

Register now: https://www.snowflake.com/summit/

Featured: Why You Need a Semantic Model

The Snowflake engineering blog published a foundational piece on why every data team needs a semantic model. As AI agents and natural language interfaces become the primary way users interact with data, semantic models become the essential bridge between raw schemas and business meaning - ensuring metrics are consistent, queries are correct, and governance is maintained.

Cortex AI & Agents

The Agentic Enterprise Control Plane

Snowflake's vision for the Agentic Enterprise Control Plane - a unified layer for orchestrating, governing, and monitoring AI agents across the enterprise. Agents operate within defined guardrails, coordinated by a central control plane built on trusted data.

Ontology on Snowflake: From Architecture to Deployment

A comprehensive guide on building operational ontologies in Snowflake - graph reasoning without a graph database. Define canonical entities, relationships, and constraints, then deploy them with a dedicated Cortex Code skill for automated ontology stack generation.

Resources:

Batch Cortex Search

Batch Cortex Search is now in public preview. The CORTEX_SEARCH_BATCH table function enables high-throughput offline search for entity resolution, deduplication, and clustering workloads.

Resources:

Cortex AI Services Cost Analyzer

Open-source tool for analyzing and visualizing costs across Cortex AI services - search, agents, analyst, and AI functions.

GitHub: https://github.com/alexfrancisross/cortex-ai-services-cost-analyzer

Agents Observe

Framework for observing and monitoring AI agent behavior, tool usage, and performance in production.

GitHub: https://github.com/simple10/agents-observe

Cortex Code: Skills, Subagents & Community

Supercharge Cortex Code CLI: Skills, Subagents, Hooks, and MCP

A practical guide covering the full extensibility surface of Cortex Code CLI - custom skills, subagent delegation, lifecycle hooks, and MCP server integration.

Build Custom Skills in Snowflake Cortex Code CLI

Step-by-step guide to creating custom skills that extend Cortex Code's capabilities for your team's specific workflows.

Demystifying Snowflake Cortex Code (CoCo)

Comprehensive overview of Cortex Code's architecture, capabilities, and how it fits into data engineering workflows.

Build a Production-Ready dbt Project Using Cortex Code CLI

End-to-end walkthrough of building a production-grade dbt project entirely driven by Cortex Code CLI.

Cortex Code DCM + dbt Demo

Demo project showcasing Database Change Management (DCM) and dbt workflows orchestrated by Cortex Code.

GitHub: https://github.com/errahulka/snowflake-coco-dcm-dbt-demo

Agentic Migration Workshop Skill

Cortex Code skill for guiding database migrations to Snowflake from Oracle, Teradata, Redshift, and SQL Server.

GitHub: https://github.com/Snowflake-Solutions/cortex-code-skills/tree/main/skills/agentic-migration-workshop

Claude Skill for Cortex Code

Community-built skill integrating Claude capabilities into Cortex Code workflows.

GitHub: https://github.com/sfc-gh-tjia/claude_skill_cortexcode

Build with CoCo

Collection of examples and patterns for building applications with Cortex Code.

GitHub: https://github.com/vinodhini-sd/build-with-coco

How Agentic Engineering Changed the Way I Build Streamlit Apps

Streamlit blog on how agentic AI workflows are transforming the Streamlit development experience.

Snowflake Platform Updates

CTAS Support for Unity Catalog External Volumes (Mar 31)

CREATE TABLE AS SELECT (CTAS) is now supported for Iceberg tables backed by Unity Catalog external volumes, simplifying table creation workflows in multi-engine lakehouse environments.

Snowflake Directory Imports

Snowflake now supports directory imports, allowing you to import an entire directory of files as a module in Python stored procedures and UDFs.

Network Policy Advisor

New Network Policy Advisor helps you understand your network access patterns and build effective network policies with data-driven recommendations.

Documentation: https://docs.snowflake.com/en/user-guide/network-policy-advisor

RBAC Hardening Blueprint

Blueprint Manager template for hardening role-based access control (RBAC) configurations in Snowflake.

GitHub: https://github.com/Snowflake-Labs/blueprint-manager/tree/main/blueprints/rbac-hardening

Semantic Views & Intelligence

Why You Need a Semantic Model

The engineering blog's deep dive into why semantic models are essential - bridging raw data and business meaning, powering consistent metrics, and enabling trustworthy natural language queries.

Snowflake-Trino SQL Translator

Open-source SQL translator for converting between Snowflake and Trino SQL dialects - useful for multi-engine lakehouse environments.

GitHub: https://github.com/Angel-Asensio/snowflake-trino-translator

Streamlit

Migrate Tableau to Streamlit in Minutes Using AI

Guide on using prompt-driven AI to migrate Tableau dashboards to Streamlit applications in minutes rather than weeks.

Apache Iceberg & Open Lakehouse

Apache Data Lakehouse Weekly (Mar 16 - Apr 2)

Alex Merced's roundup of the latest across the Apache data lakehouse ecosystem - Iceberg, Polaris, and open table format developments.

Data Sharing & Open Table Formats

Snowflake's vision for data sharing with open table formats enables seamless data exchange while maintaining interoperability across engines and platforms.

Machine Learning & AI Guides

Defect Detection Using Distributed PyTorch with Snowflake Notebooks

Build a defect detection model using distributed PyTorch training directly in Snowflake Notebooks - no external infrastructure needed.

Guide: https://www.snowflake.com/en/developers/guides/defect-detection-using-distributed-pytorch-with-snowflake-notebooks/

Getting Started with Snowpipe Streaming on AWS MSK

Step-by-step guide for setting up Snowpipe Streaming with Amazon Managed Streaming for Apache Kafka (MSK).

Guide: https://www.snowflake.com/en/developers/guides/getting-started-with-snowpipe-streaming-aws-msk/

Deploying Gemma 4 Multimodal on Snowflake SPCS

Deploy Gemma 4 multimodal model on Snowpark Container Services for text and image inference using llama.cpp.

Gemma 4 on Ollama

Gemma 4 is now available on Ollama for local inference.

Run it: https://ollama.com/library/gemma4

Validating a Data Platform Migration with an AI Coding Agent

Real-world account of using an AI coding agent to validate a data platform migration - covering test generation, assertion patterns, and confidence building.

dbt LLM Evals

Framework from Paradime for evaluating LLM-generated dbt models - measuring correctness, efficiency, and style compliance.

GitHub: https://github.com/paradime-io/dbt-llm-evals

Snowflake Labs Repositories

Repository	Description
unified-data-stack-for-incident-management	Reference architecture for building a unified incident management data stack on Snowflake
MCP-HOL-BUILD-2025	Hands-on lab for building MCP integrations with Snowflake
blueprint-manager (RBAC Hardening)	RBAC hardening blueprint for Snowflake security best practices
sfquickstarts: Batch Cortex Search	Setup SQL for the Batch Cortex Search quickstart

From the Community: Medium & Blogs

Article	Description
Ontology on Snowflake: Architecture to Deployment	End-to-end guide for building and deploying ontology stacks on Snowflake with Cortex Code
Operational Ontologies in Snowflake	Graph reasoning without a graph database using Snowflake's semantic and relational layers
Migrate Tableau to Streamlit in Minutes	Use prompt-driven AI to convert Tableau dashboards into Streamlit apps
Snowflake Supports Directory Imports	Import entire directories as Python modules in stored procedures and UDFs
Batch Cortex Search: Hybrid Search for High-Throughput Workloads	Engineering deep dive into the batch search architecture
Build a Production-Ready dbt Project with Cortex Code	End-to-end production dbt project driven by Cortex Code CLI
Build Custom Skills in Cortex Code CLI	Step-by-step custom skill creation guide
Demystifying Snowflake Cortex Code	Comprehensive CoCo architecture and capabilities overview
Deploying Gemma 4 Multimodal on Snowflake SPCS	Text and image inference with llama.cpp on Snowpark Container Services
How Agentic Engineering Changed the Way I Build Streamlit Apps	How agentic AI is transforming Streamlit development
Validating a Data Platform Migration with an AI Coding Agent	Using AI agents for migration validation and testing
Supercharge Cortex Code CLI: Skills, Subagents, Hooks, and MCP	Practical guide to the full CoCo extensibility surface

MCP & Agent Tooling

Tool	Description
Coalesce MCP	MCP integration for the Coalesce data transformation platform
Sidemantic	Semantic search MCP server for local document collections
MCP HOL Build (Snowflake Labs)	Hands-on lab for building MCP integrations with Snowflake
AutoSkills	Auto-generate AI agent skills from documentation and examples
METATRON	Meta-agent framework for orchestrating specialized AI agents

AI Research & Models

Gemma 4 Multimodal

Google's Gemma 4 multimodal model is now available - handling both text and image understanding. Run it locally via Ollama or deploy on Snowflake SPCS.

Ollama: https://ollama.com/library/gemma4
Snowflake SPCS Deployment Guide: https://adrianleexinhan.medium.com/deploying-gemma-4-multimodal-on-snowflake-text-and-image-inference-with-llama-cpp-and-spcs-76eec0a6d444

Athena Core

Open-source AI agent core from Athena AI Lab for building research and reasoning agents.

GitHub: https://github.com/Athena-AI-Lab/athena-core

A-Evolve

Self-evolving AI agent framework from A-EVO Lab that learns and adapts from its own execution history.

GitHub: https://github.com/A-EVO-Lab/a-evolve

OpenClaude

Open-source Claude-compatible interface for running and testing AI agent workflows.

GitHub: https://github.com/Gitlawb/openclaude

Duck Lineage

Data lineage tracking for DuckDB pipelines - visualize how data flows through your transformations.

GitHub: https://github.com/ilum-cloud/duck_lineage/

PaddleOCR

Baidu's state-of-the-art OCR toolkit supporting 80+ languages with high accuracy.

GitHub: https://github.com/PADDLEPADDLE/PADDLEOCR

Developer Tools

Browsers, Automation & Productivity

Tool	Description
Post4U	Schedule and automate social media posts across platforms
OwnScribe	Self-hosted writing and note-taking application
OpenScreen	Open-source screen sharing and recording tool
World Monitor CopilotKit	Global monitoring dashboard built with CopilotKit
Happier	Developer happiness and productivity toolkit
OfficeCLI	Command-line interface for office productivity tools
Ste.gg	Modern personal website and portfolio builder

Code, Icons & Design

Tool	Description
Developer Icons	Curated collection of developer and technology brand icons
FlipOff	CLI tool for quick file and project scaffolding

Documents, Media & Data

Tool	Description
Exifa	EXIF metadata analyzer and viewer for images
PDFx	PDF extraction and processing tool
Podcats	Podcast cataloging and management tool
Video Explainer	AI-powered video analysis and explanation tool
Call.md	Markdown-formatted call transcription and notes
GhGrab	GitHub repository downloader and archiver

Security & OSINT

Tool	Description
Checkov	Static analysis for infrastructure-as-code security scanning (Terraform, CloudFormation, Kubernetes, etc.)
QGIS OSINT	OSINT plugin for QGIS geographic information system

Chat & Communication

Tool	Description
Atomic Chat	AI-powered chat application with multi-model support
Strix	Lightweight messaging and communication framework
Mick	Conversational AI assistant framework
Frosty	AI chatbot built with Snowflake integration
Wardrowbe	AI-powered wardrobe management and outfit recommendation
Bluesky Bot Tutorial	Official AT Protocol guide for building Bluesky bots

Health & Awareness

Sytse Sijbrandij (GitLab co-founder) shares his cancer journey - an important read on health awareness for the tech community.

Read more: https://sytse.com/cancer/

ROI of Generative & Agentic AI

Snowflake published extensive research demonstrating the radical ROI organizations are seeing from generative and agentic AI implementations.

Resources:

Data Sharing & Open Table Formats

Snowflake's vision for data sharing with open table formats enables seamless data exchange while maintaining interoperability across engines and platforms.

SNACK-AI: The 2026 Pattern

As we look forward in 2026, SNACK-AI is the architectural pattern for powering AI applications at any scale - from mobile phones and robots to massive enterprise clusters.

SNACK-AI integrates: