Syeda Sara Fathima’s Post

🧠 It used to be that when running a large language/reasoning model, you had to trade off privacy and locality by using the provider's hosted service. With the arrival of #DeepSeek R1, these models are now small and efficient enough to run yourself, on commodity GPUs. ❄️ What better place to run this model, than close to your data in Snowflake, with full governance and auditing. The model runs safely in a container, with you fully in control of any egress. And you don't have to trade off performance: run the 32B distill with its default context window on all 4 A10 GPUs via tensor parallelism. It's a fast and secure setup, that only takes minutes to bring up. 👉 Try it out below!

DeepSeek deserves a closer look; now you can do so, on Snowflake. #deepseek #snowflake #snowparkcontainerservices #LLMs

🚀 Finally got DeepSeek-R1 running in Snowflake Container Services and I couldn't be more excited to share this one! (Probably the coffee talking, but hey! ☕️) After some seriously deep diving (and yes, a concerning amount of caffeine), I've put together a complete guide on deploying DeepSeek's incredible 671B parameter model right inside Snowflake's secure environment. What's got me really jazzed about this setup? Imagine having a 671B parameter model's reasoning capabilities distilled into 32B that you can actually run in production! We're talking: - Enterprise-grade security that just works - Zero infrastructure headaches (goodbye control plane management!) - Direct data access right where you need it - Complete control over your environment The best part? I've included everything from Dockerfile setups to UI implementation in the guide. Whether you're diving into LLMs for the first time or you're a seasoned pro, this setup makes deploying advanced AI models smoother than ever. 👉 Want the full step-by-step guide? Check it out here:https://lnkd.in/gh-bD-jQ Drop me a message or connect if you want to geek out about containers and AI! #Snowflake #AI #DeepSeekAI #TechLife #DeveloperAdvocate #Innovation #LLM #ContainerServices

Hey data enthusiasts! ❄️ My colleague James Cha-Earley has written an incredible step-by-step guide on how you can run DeepSeek-R1 in a secure enterprise-ready environment in Snowflake, check it out 👇

Struggling to get your GenAI proof of concept into production? Come along next week in London to hear me talk about how Databricks can empower you to productionise your GenAI use cases cost-effectively whilst using your own data. #databricks #genai #llms #mosaicai  #dataintelligenceday

#snowflake_advocate FROSTY_FRIDAY() is a series of weekly Snowflake challenges to help users develop their skills in the platform, this week we're using Cortex ML! try it here: https://lnkd.in/edyuGBjE

Are you ready to take your document classification game to the next level with Snowpark? This blog dives deep into the realm of Snowflake ML, demonstrating how you can leverage Snowpark for document classification. Learn how to harness the power of Snowflake, combined with machine learning capabilities, to streamline document organization and classification processes like never before! 🗃️ #SnowparkML #Snowflake #MachineLearning #TechBlog #phData #MLE

Had an incredible experience at the Snowflake BUILD Demo featuring Cortex Analyst! Huge shoutout to Jason Holt for breaking down complex concepts into bite-sized, easy-to-understand pieces. And of course, a big thanks to the awesome Snowflake 🏂 dev team! The best part? I managed to create this app while running the event simultaneously. So, trust me when I say, if I can do it, YOU can too! One more thing...I'm taking my Snowflake journey to Medium. Follow along as I dive deep into all things Snowflake ❄️. Let's learn and grow together! #Snowflake #BUILD #CortexAnalyst #AI

Dr Andrew Ng is speaking at a Snowflake conference about the recent trends surrounding AI agents, which have been gaining significant attention, as well as the new tech stack supporting them. Although it’s a Snowflake event, his talk is quite general and not limited to Snowflake. This makes it a valuable video for anyone currently working on LLM PoCs to gain insights into future trends. https://lnkd.in/g2ygTB2K

Join your Looker developer community and explore the innovative possibilities Vertex AI can bring to Looker and Looker Studio data experiences. Get your hands on the technical possibilities with Vertex AI’s Generative AI and ML models, Looker’s Extensions, Actions, Embedding, API, open-source tools, Looker Studio Community Connectors and Community Visualizations, and more. Aim for the "Best Hack" Award and learn something new along the way! Official Rules #GoogleCloud #VertexAI #Looker #AI