Snowflake’s Post

We’ve renewed our Premier Partnership with Snowflake. Since 2017, we’ve been building with Snowflake as their first local partner. That matters. We’ve seen the platform evolve from early adoption to becoming the backbone of modern data stacks across Europe. And we’ve been in the trenches with our clients every step of the way; designing, building, fixing, and scaling what actually works. This renewal isn’t a milestone. It’s continuity. We stay close to the technology. We stay close to our clients. And we keep pushing for systems that deliver real business impact. On to the next phase.

It’s the time of year when I suspect both Snowflake and Databricks have started hoarding releases to announce at their summits in June. Looking forward to it.

The biggest question re: the future of Databricks and Snowflake isn’t valuation, market share, or feature set. It’s whether we run out of lake-* product names or snow puns first.

Every collections company claims some use of AI now. It's becoming table stakes. As a result, it's becoming more important for lenders to ask: What does the underlying infrastructure look like and who built it? Be weary of vendors with thin answers. What data is feeding their models? Most agencies bolt point solutions onto weaker datasets. January's models learn from 1B+ borrower interactions. Tyler Gillies nails it, MLOps is what turns a good model into one that actually improves every month. Infrastructure from experiment to production to continuous improvement. January doesn't outperform because we use AI. We outperform because we built the data engine that makes AI work in collections years ago.

Here's what we learned at FabCon in Atlanta. 8,500 attendees. Real companies. Real results. Six takeaways every CTO should know. 1. Big companies aren't using Fabric for dashboards, they're running full operations through it. 2. Lumel Planning Writeback is now a first-class Fabric workload, included in your license at no additional cost. 3. OneLake and Snowflake both support Apache Iceberg, meaning you can run bi-directional integration with Fabric, Databricks, and Snowflake without copying a single byte of data. 4. Microsoft Purview and Entra ID classify data at the source, and that classification follows it everywhere, including into AI agents. 5. AI agents are useful, but getting one to take the right action repeatedly is where most deployments break down. 6. SQL Server 2025 ships with native AI capabilities and meaningfully better JSON handling for organizations still running on-premise workloads. The full breakdown drops Wednesday. If any of these hit close to home, the link to schedule a conversation with our team is in the comments.

🚀 𝗛𝘂𝗴𝗲 𝗱𝗮𝘆 - 𝗬𝘂𝗸𝗶 𝗻𝗼𝘄 𝘀𝘂𝗽𝗽𝗼𝗿𝘁𝘀 𝗚𝗼𝗼𝗴𝗹𝗲 𝗕𝗶𝗴𝗤𝘂𝗲𝗿𝘆 https://lnkd.in/dX6bpUZU 🚀 When we started building Yuki, - we focused on Snowflake. As we grew, we started hearing something new from customers: they didn’t just want optimization - they wanted flexibility. More teams today are running hybrid data stacks. More companies are actively avoiding vendor lock-in. The reality is simple: teams don’t want to commit to one data platform anymore. That’s why this milestone matters - Yuki now supports Google BigQuery! Huge shoutout to our entire engineering team - this was a massive effort, and I couldn’t be more proud! This is an important step toward something bigger: helping teams avoid vendor lock-in - while keeping their data infrastructure efficient, predictable, and scalable. And we’re just getting started.

Moving data just to run ML slows everything down. Snowflake ML lets you build, deploy, and serve models directly where your data lives with sub-100ms inference.  It's faster, simpler, and with no extra infrastructure.  See for yourself: https://lnkd.in/eevruN8A

Moving data just to run ML slows everything down. Snowflake ML lets you build, deploy, and serve models directly where your data lives with sub-100ms inference.  It's faster, simpler, and with no extra infrastructure.  See for yourself: https://lnkd.in/gB4_s4rK

Getting started with Snowflake Feature Store?  This guide walks through building entities, feature views, and datasets using Snowflake Notebooks—laying the groundwork for scalable ML pipelines.  Check it out ⬇️

Moving data just to run ML slows everything down. Snowflake ML lets you build, deploy, and serve models directly where your data lives with sub-100ms inference.  It's faster, simpler, and with no extra infrastructure.  See for yourself: https://lnkd.in/gpiwNZ5e