Filipe Gracio, PhD

"Build a Data Lakehouse with Apache Iceberg, Polaris, Trino & MinIO" Enjoyed reading this two-part blog post series by Gilles Philippart, teaching how to run a Data Lakehouse locally, using open-source technologies: - MinIO for object storage - Apache Iceberg as the open table format - Apache Polaris as the metadata catalog - Trino as the query engine. The second part adds Apache Kafka and Flink for streaming data ingest. I also learned about JR, a CLI utility for creating randomized Kafka test data. Everything running locally on your machine via Docker Compose, which is great for learning and experimenting. Really neat, definitely recommend to check it out! 👉 https://lnkd.in/e73fSnPS

201

13 Comments

🔊 Audio just became AI-native in Snowflake. AI_TRANSCRIBE is now available in Cortex — meaning you can transcribe audio files, detect speakers, and get word-level timestamps without managing any ML infra. Combine it with AI_SENTIMENT or AI_CLASSIFY and suddenly every call, podcast, or voice note becomes searchable, labelable, and insight-ready. I made a new Quickstart that walks through exactly how to set this up — including combining transcription with text and image analysis: 👉 https://lnkd.in/gyxeDdPi Can't wait to see how folks use this in production 🔥 #Snowflake #CortexAI #AITranscribe #MultimodalAI #GenAI #VoiceAI

112

5 Comments

ICYMI - This Week in AI (Week of June 8th) is now live! ⚡ Here’s your 5-minute pulse check on what actually moved the needle in AI this week: → Snowflake's New Brain: Data Giant Unveils "Intelligence Cloud" and Acquires Crunchy Data → OpenAI Sounds the Alarm: Uncovers Covert AI-Powered Influence Operations → AI Coding Assistant ‘Cursor’ Nabs Nearly $1 Billion, Reaches Staggering $9.9B Valuation → Google DeepMind is Building an AI to Run Your Inbox → From Text to Emotion: ElevenLabs Launches Hyper-Realistic v3 Voice AI → Stability AI Drops Stable Diffusion 3 Medium: Better, Faster, and More Photorealistic Images If you’re building, investing, or just AI-curious, these are the signals worth tracking. PS: I do the reading, so you don’t have to. No jargon, no fluff, just sharp, relevant AI updates 💙 📖 Read the full newsletter: https://lnkd.in/dzWXN2DT Co-author: Harini Anand + AI 🤖 ------------ ♻️ Share this with your network to keep them in the loop 👩‍💻 Follow me (Aishwarya Srinivasan) for weekly AI insights, product breakdowns, and industry signals that matter

148

36 Comments

DeepSeek's R1 model is generating a lot of excitement, and for good reason. We've been in an era of "bigger is better,” with a handful of AI companies dominating the landscape. R1 shows us that smart optimization and clever training strategies can deliver exceptional results with less. That’s why we’re excited to bring DeepSeek’s models directly to organizations’ data where it already resides in the Snowflake platform, so that customers can use R1 while still benefiting from Snowflake’s robust security and governance guardrails. As a next step, our Snowflake AI Research team plans to enhance R1 to further reduce the inference costs, so customers can achieve even greater cost-performance efficiencies with R1 and expedite the delivery of generative AI apps! Check out today’s blog post to learn more.  https://lnkd.in/gn2upTKg

147

3 Comments