:probabl.

Earlier this week, ✨ Marie Sacksick interviewed Merel Theisen, Tech Lead of Kedro at QuantumBlack, AI by McKinsey, about building the open source tools enabling enterprise data science teams to scale their impact. 🗣️ Marie and Merel talked about what it really takes for enterprise data science teams to move from experimentation to production: standardized pipelines, composable open source tools, and the kind of rigor that turns individual work into institutional practice. Merel also shared great insights on how Kedro complements tools like scikit-learn and #Skore, and what a healthy open source data science ecosystem looks like to her. 👉 Read or listen to the interview on our blog: https://lnkd.in/ewCGfCCH #AI #MachineLearning #ScikitLearn #Python #EnterpriseAI

I discovered Kedro a few years ago thanks to my former colleague Pierre Godard at CybelAngel. It solved our problems of rewriting notebooks to python script before releasing a model in production. That’s why I’m excited to share this conversation with Merel Theisen, Tech Lead of Kedro at QuantumBlack, AI by McKinsey. We talked about what it really takes for enterprise data science teams to move from experimentation to production: standardized pipelines, composable open source tools, and the kind of rigor that turns individual work into institutional practice. Merel shares great insights on how Kedro complements tools like scikit-learn and Skore, and what a healthy open source data science ecosystem looks like. Thanks Cailean Osborne for the facilitation! 🗞️ https://lnkd.in/eurEQTVQ

Skills in machine learning and AI are reshaping salaries and hiring decisions – and the data proves it. Research by Dr. Fabian Stephany at the University of Oxford shows that professionals with machine learning skills earn up to 40% more than their peers who don't have such skills. Those who can back up their skills with certifications have a 15% higher chance of being invited to interviews. Last week, David Arturo Amor Quiroz, who leads our Skolar training and certification program, sat down with Fabian to dig into what these findings mean for professionals and recruiters alike. Read or listen to the interview on: 🔗 Our blog: https://lnkd.in/dbR4_kQg 🔗 Substack: https://lnkd.in/dJ_dgmZG 🔗 Medium: https://lnkd.in/dM__AZjx #AI #MachineLearning #ScikitLearn #Python #FutureofWork

Experiment tracking? The tools exist. Knowing which experiment actually matter? That's the hard part. If you've used Neptune.ai, MLFlow, Weights&Biases, or Comet, you know the drill. Hundreds of runs logged. But which one moves the needle for the business? That's exactly why we are building Skore. Join us on April 9 at 4 PM (Paris time) for a live webinar with Guillaume Lemaitre from Probabl, the company behind scikit-learn, to see how Skore helps you: → Structure and track your ML experiments → Showcase your best models → Report results to stakeholders Skore works alongside your existing tools. No rip-and-replace. Don't just track change, create an advantage. 🔗 Register now: https://lnkd.in/ewbDiJJD

At :probabl. we firmly believe that empirical data about the business models of open source companies is invaluable for building challenger companies upon the foundation of world-class open source projects – like scikit-learn. For this reason, we’re proud to support this essential research by The Linux Foundation alongside our partners Serena, Mistral AI, Amazon Web Services (AWS), Red Hat, Futurewei Technologies, Inc., the Cloud Native Computing Foundation (CNCF), the Commercial Open Source Startup Alliance (COSSA), MistyWest, and Y Combinator. 👋 Does your company build and/or commercialize open source software? Help us establish a baseline for the global open source community by taking this short survey: https://lnkd.in/d_23c2Cs

Earlier this week, David Arturo Amor Quiroz interviewed Dr. Fabian Stephany, a world-leading expert on AI and the future of work at the University of Oxford, about the value-add of certifying your ML and AI skills. Two findings stood out to us: 1️⃣ By having general AI-related skills, professionals can earn 21% more than those who don’t have such skills. With ML skills, that boost reaches up to 40%. 2️⃣ As generative AI makes it increasingly hard for recruiters to identify candidates with genuine skills, having certified ML and AI skills in your resume increases the likelihood of landing an interview invitation by 15%. To learn more, read or listen to Arturo's interview with Fabian on: 🔗 Our blog: https://lnkd.in/dbR4_kQg 🔗 Our Substack: https://lnkd.in/dJ_dgmZG 🔗 Our Medium: https://lnkd.in/dM__AZjx #AI #MachineLearning #ScikitLearn #Python #Skills #FutureofWork

At :probabl. we are committed to democratizing AI, not only by stewarding scikit-learn but also by developing free educational materials for learning machine learning with Python and our official Skolar certifications. This week, I stepped away from my usual work on Skolar to speak to Dr. Fabian Stephany from the Oxford Internet Institute, University of Oxford, one of the world’s leading researchers on AI and the future of work, to learn about what he and his research team have found about the value of AI skills in today’s job market. Two findings from Fabian’s SkillScale Project stood out to me in particular: 1️⃣ By having general AI-related skills, professionals can earn 21% more than those who don’t have such skills. With machine learning skills, that boost reaches up to 40%. 2️⃣ As generative AI makes it increasingly hard for recruiters to identify candidates with genuine skills, having 𝐜𝐞𝐫𝐭𝐢𝐟𝐢𝐞𝐝 𝐦𝐚𝐜𝐡𝐢𝐧𝐞 𝐥𝐞𝐚𝐫𝐧𝐢𝐧𝐠 𝐬𝐤𝐢𝐥𝐥𝐬 in your resume increases the likelihood of landing an interview invitation by 15%. To learn more, read or listen to my interview with Fabian on: 🔗 Probabl blog: https://lnkd.in/eVJfYEYE 🔗 Substack: https://lnkd.in/eF-QHfrA 🔗 Medium: https://lnkd.in/e459fMzQ Huge thanks to Fabian for the fascinating conversation and to all my colleagues and peers at :probabl., Inria, and the wider scikit-learn community for helping to democratize AI, starting with providing free access to state-of-the-art educational resources and open source software.

This week, our CSO Gael Varoquaux interviewed NVIDIA's Dr. Andy R. Terrel (he/him) about the engineering efforts that have enabled GPU acceleration in scikit-learn and the benefits for the millions of data scientists using scikit-learn in research labs and enterprises across the world. In the interview, Gaël and Andy discussed the work by the scikit-learn community to adopt the Python array API and the wins for the millions of data scientists. who use scikit-learn in their data science workflows across diverse research and industry contexts. At the end, Andy shares how he'd pitch scikit-learn to his boss, Jensen Huang, and his answer was golden. To learn more, read or listen to the interview on: 🔗 Our website: https://lnkd.in/dxxtTDHP 🔗 Our Substack: https://lnkd.in/dRB2VkSG 🔗 Our Medium: https://lnkd.in/d6keei3b #ScikitLearn #MachineLearning #DataScience #Python #OpenSource

We just wrapped up our very first TabICL sprint! The TabICL team (Jingang Qu, David Holzmüller, Gael Varoquaux and I) joined forces with scikit-learn developers at :probabl. Olivier Grisel and Loïc Estève to ensure we follow open-source best practices and gather technical feedback. Highlights ✅ TabICL now has a website! (see link in comment) ✅ New experimental features protyped: SHAP computation, density estimation, data generation, anomaly detection. ✅ Verified scikit-learn compliance. Many thanks Loïc and Olivier for your time and expertise, it was great to spend this time with you! #OpenSource #TabICL #TabularFoundationModel

Our CSO Gael Varoquaux was on stage at NVIDIA GTC today discussing what we need to do to accelerate open science. The panel covered the full gamut of what’s needed to speed up core scientific Python libraries, from the benefits of standardizing abstraction layers to the need for greater resource allocation. Gaël shared how the scikit-learn maintainers have been hard at work to enable GPU acceleration, benefiting millions of data scientists who use scikit-learn every day. To learn more about what’s going on in scikit-learn, read or listen to Gaël’s interview with Dr. Andy R. Terrel (he/him) published today: 🔗 Probabl blog: https://lnkd.in/dxxtTDHP 🔗 Substack: https://lnkd.in/dRB2VkSG 🔗 Medium: https://lnkd.in/d6keei3b Thanks to Leo Fang , Ianna Osborne, Travis Oliphant, Katrina Riehl, and the NVIDIA team for the fascinating panel. Onwards and openwards – at GPU speed 🚀