Featureform

? What does "real-time" really mean in the context of machine learning? Our latest blog post on "Breaking Down "Real-Time" Machine Learning Systems" explores these topics and more. It dissects real-time ML systems into their various components, such as online inference, low-latency serving, and stream processing. This article will help you navigate the complexities of deploying a real-time ML system. Whether you're architecting it from scratch or working on an existing system, demystifying the often-overloaded term "real-time machine learning" will help you build better platforms. ?? Check it out: https://buff.ly/4ayIE4P

?? Ready to elevate your data infrastructure to meet petabyte-scale demands? Join our founder Simba Khadder on Tuesday, December 3rd at 8 AM PT for "ML Feature Lakehouse: Building Petabyte-Scale Data Pipelines with Iceberg!" The term "Feature Store" might sound like just a place to store features, but in reality, it’s a powerful system for defining, managing, and deploying large-scale data pipelines. This session will simplify feature stores by breaking down the three main types and showing how they fit into an ML ecosystem. We’ll explore how feature stores enable data scientists to build, manage, and scale their pipelines, even at petabyte levels, while handling streaming data and ensuring versioning and lineage. We’ll also look under the hood to see how Featureform achieves this scale using Apache Iceberg so you leave with actionable insights to improve your ML platforms and projects.

?? Our team had a fantastic time at MLOps World & Generative AI World last week! ?? Many thanks to Faraz Thambi and Tina Aprile, CMP, for making this conference happen. We'd also like to thank Modal (especially Alec Powell and Charles Frye) and Jared Zoneraich, founder of PromptLayer, for being amazing happy hour co-hosts! ??

?? Join our founder,?Simba Khadder,?on November 12th at 8 AM PT, where he will explore the potential of RAG beyond its current use cases. Today's standard approach to RAG—embedding a user query, performing a nearest neighbor search on arbitrarily chunked text via a vector database, and fitting it into a prompt—is far from optimal. While it works, this technique misses the opportunity to truly optimize how much relevant information we can fit into the context window. The real challenge is squeezing the most useful data into the prompt, and there’s a wealth of research and techniques available that can significantly improve this process. In this one-hour session, we’ll explore why the current approach is limiting and how we can leverage smarter strategies to optimize context, improve retrieval, and unlock the full potential of RAG. Toward the end of the session, there will also be time for Q&A.

?? Looking for somewhere to unwind after day one of MLOps World in Austin? Join us Thursday at 5 PM CT for a happy hour with Modal and PromptLayer at Estancia Brazilian Steakhouse! ? RSVP here to attend: https://lu.ma/pu3ivvzr See you then ??

?? Will you be in Austin for MLOps World / Generative AI World this Thursday and Friday? Come visit our booth! Our CEO Simba Khadder will also be hosting a couple of talks during the conference: ?? On Thursday, November 7th, at 1:35 PM CT, Simba is hosting a lightning talk: "We’re Doing RAG All Wrong—and How We Can Do So Much Better" ?? On Friday, November 8th, at 12:35 PM CT, Simba is hosting a speaker's corner to discuss how Data Scientists can build Petabyte-Scale Pipelines with Iceberg. See you then!

?? Check out our founder Simba Khadder as he chats with Omar Shanti and Matt Paige of HatchWorks AI about the connections between generative AI and traditional ML in the latest Talking AI Podcast, "Generative AI is Not an Island: ML’s Core Principles." They also discuss how LLMs build off traditional ML fundamentals, the evolution of RAG, and what it takes to scale AI in real-world environments. ?? Listen to the full podcast here: https://buff.ly/4eSvhz5

Did you miss Simba Khadder's talk at the Data Engineering for AI/ML conference? Don't worry, we got it right here for you ?? ?? Watch the full seminar here: https://buff.ly/489eK7j ?? Huge shout out to Demetrios Brinkmann and the MLOps Community for hosting such a great conference!

The term "Feature Store" often conjures a simplistic idea of a storage place for features. However, in reality, feature stores are powerful frameworks and orchestrators for defining, managing, and deploying data pipelines at scale. Join Simba Khadder, founder and CEO of Featureform, on October 15th at 8 A.M. PT as he demystifies feature stores, outlining the three distinct types and their roles within a broader ML ecosystem. We’ll explore how feature stores empower data scientists to build and manage their own data pipelines, even at petabyte scale, while efficiently processing streaming data, and maintaining versioning and lineage. At the end of the session, there will also be time for Q&A. We hope to see you there!

?? We had an awesome happy hour last week! Big shoutout to the Modal team, including Alec Powell Powell and Charles Frye, for co-sponsoring last week's happy hour with us. We had a blast with you guys! ??