Outerbounds

🚀 Tired of customizing Metaflow workflows via code changes? Our friends at Netflix have introduced Configurable Metaflow! This feature allows you to define variants of your workflows using config files, facilitating experimentation with different data, parameters, or algorithms without altering your core code. Read more: https://hubs.la/Q030nDpk0

At the Gen AI World Summit in Austin, our own Savin Goyal tackles one of the biggest pain points in ML workflows: managing dependencies and ensuring reproducibility at scale. Drawing on his experience at Netflix and Outerbounds, Savin walks through why dependency mismatches, environment inconsistencies, and scaling challenges can slow innovation—and how to fix them. Savin introduces how Metaflow addresses these issues with features like isolated virtual environments, cross-platform reproducibility, and rapid image generation. The result? Data scientists can iterate faster and focus on solving problems instead of managing infrastructure. If you're building AI infrastructure or supporting data science teams, this talk provides real lessons to improve productivity and scalability: https://hubs.la/Q030hgKn0

If you're in the Bay Area, join us tomorrow evening at GitHub's SF office! Our Eddie Mattia will be discussing fine-tuning of LLMs using Metaflow, covering models ranging from 1 billion to 405 billion parameters 💪 RSVP @ https://hubs.la/Q02-YS890

Over the past years, we have worked with many organizations leveraging (or would like to leverage) hybrid compute environments, including their existing HPC clusters, orchestrated with Slurm. Our new @slurm decorator feels magical: with just a few lines of code, you can run workloads in your Slurm clusters and burst into the cloud seamlessly while benefiting from all the features of Metaflow. The best part? It's now open-source! Read more @ https://lnkd.in/gGgdfMRb

Join us next Tuesday evening, December 10, at GitHub's SF office for an exciting event featuring our very own Eddie Mattia where he'll discuss cost-efficient fine-tuning of LLMs using Metaflow, covering models ranging from 1 billion to 405 billion parameters. Eddie will demonstrate how Metaflow enables seamless integration of fine-tuning into workflows, utilizing popular tools and cloud-based compute resources. RSVP @ https://hubs.la/Q02-44-40

💫 I had a wide ranging conversation this week with Alexander Filipchik for Outerbounds about how to turn machine learning and AI into actual engineering disciplines. Topics below 👇 Watch here: https://lnkd.in/gH896Uus

The new @slurm integration in Metaflow is a big hit at #supercomputing #sc24 😊 If you are in Atlanta, come to our booth 516 to see a demo in person. The second best option is to read our new blog post about it If you are truly impatient, just "pip install metaflow-slurm" https://lnkd.in/gkMi8355

Going live in 10 with Alexander Filipchik for Outerbounds to chat about ML and AI engineering at Cloudkitchens 🍳 https://lnkd.in/gSbvTWiE

How are ML and AI are evolving into actual robust engineering disciplines? ....Or are they? 🤔 Join me and Alexander Filipchik(Head of Infrastructure, City Storage Systems AKA CloudKitchens) for an Outerbounds fireside chat tomorrow where we'll discuss this and more 💫 https://lnkd.in/gQ7P3_jT