Hi everyone! Tomorrow at 13:00 CET I’ll be speaking at DataZen about using Kedro to bring structure and reproducibility to messy AI projects. I’ll walk through problems from real projects and stick around after the talk for Q&A — happy to connect and chat. https://datazen.top/8a4c UPD: for those interested in the recording: https://www.youtube.com/live/0979QPTCGtg?si=UrWr8BxlVMziVtjx

pretty barebones, written on a whim (3 commits) but interesting concept! https://github.com/okua1/juvio

https://invariantlabs.ai/blog/mcp-github-vulnerability 👀

I think we share the same view on Jinja template https://www.linkedin.com/posts/toby-mao_jinja-sqlmesh-vscode-activity-733387908401747[…]m=member_desktop&rcm=ACoAAB11enwBho3z6f4K9HCr7RLToPmxcW-lAkY

https://www.linkedin.com/posts/niels-rogge-a3b7a3127_when-people-ask-me-which-framewo[…]m=member_desktop&rcm=ACoAAAXcavUB6j26_YjUtYinmzmdjTchguHHuG4 💡

NannyML, one of the most promising ML observability solutions, is shutting down and joining Soda https://www.linkedin.com/posts/hakimelakhrass_nannyml-is-shutting-down-its-been-45-ac[…]m=member_desktop&rcm=ACoAAAXcavUB6j26_YjUtYinmzmdjTchguHHuG4

TIL that the open source Unity Catalog is now 2 products: Unity Catalog and Unity Catalog AI https://github.com/unitycatalog/unitycatalog/releases I'm a bit confused

! hugging face transformers drops support for Tensorflow https://www.linkedin.com/posts/lysandredebut_i-have-bittersweet-news-to-share-yesterd[…]m=member_desktop&rcm=ACoAAAXcavUB6j26_YjUtYinmzmdjTchguHHuG4

"the case for notebooks in prod" 👀 https://www.linkedin.com/posts/eric-riddoch_the-case-for-notebooks-in-prod-part-1-act[…]m=member_desktop&rcm=ACoAAAXcavUB6j26_YjUtYinmzmdjTchguHHuG4

We know pain too lolsob https://kubernetes.io/blog/2025/06/16/changes-to-kubernetes-slack/

New eBook Alert - "Building the Control Layer for Agentic AI with AI Gateway and MCP Servers" Learn how enterprises are deploying scalable, secure, and context-aware GenAI systems on-prem, in VPCs, or hybrid environments. A practical and technical roadmap for AI officers, Data scientists, Platform heads, MLOps engineers, CDOs, CIOs, and architects building production-grade AI infrastructure with AI Gateways and Model Context Protocol (MCP). Key takeaways are - Why AI Gateways are emerging as the central abstraction and orchestration layer in the GenAI stack How MCP allows agents to discover, authenticate, and invoke enterprise tools through machine-readable, context-aware schemas Deep dive into the architecture, protocols, and design principles shaping the next generation of enterprise AI infrastructure Access the free ungated version of the eBook here - https://gdurl.com/0RO8/download

@Max Pardoe has left the channel

https://cube.dev/docs/product/apis-integrations/queries What's new about the semantic layer? Came across this and have some more thinking

In a bit of self-promotion, wanted to share a talk I gave earlier this summer at SciPy: Much of the motivation for improving the Python data engineering experience comes from pain points seen from Kedro users (including my own experience from when I was primarily a Kedro user), and the idea that Kedro is well-positioned to be the core of the Python-native "modern data stack." It's very much a community-driven thing; thanks to so many people: @Iñigo Hidalgo @Guillaume Tauzin @Mark Druffel @datajoely @Juan Luis just to name a few. If you'd like the slides: https://deepyaman.github.io/python-is-all-you-need/ Last but not least, if you are using a combination of these technologies (e.g. Kedro + Ibis + pandera, and maybe even dlt or some stuff on the BI end), and would be open in sharing more about how you use them, would love to talk! Please shoot me a message if so. 🙏

Hey Kedro community! K 🧬 Just published a blog post about using Kedro to structure BioLM protocols: https://blog.biolm.ai/translating-biolm-protocols-into-kedro-workflows/ Over at BioLM, we are using language models to power biological research. Our platform provides models for protein structure prediction, sequence generation, and molecular design. We make frequent use of Jupyter notebooks to compose multi-step modeling workflows, often for user demonstration, but sometimes for reproducibility purposes a notebook can be a bit too "fluid" so we need a more defined workflow. Kedro feels like a natural transitional fit in such cases. We used Kedro to translate a Jupyter notebook (demonstrating antibody engineering with protein structure analysis) into a structured, reproducible pipeline. I think it worked out pretty well! Some key Kedro features we found really useful for this purpose: • Configuration management: Externalized file paths, API parameters, and target definitions into YAML files • Data catalog: Made data dependencies explicit and the dataset previews in Kedro-Viz are great • Checkpoint functionality:

kedro run --from-nodes

and

--from-outputs

is super helpful for iterative development • Kedro-Viz: Publishing the static Kedro-Viz build via GitHub Actions → GitHub Pages is a great way to share and present workflows (like this: https://biolm.github.io/biolm-kedro/) The workflow processes multiple protein targets (EGFR, PDL1, MBP, IL-7Rα) through BioLM's AI models, generating variant analysis and visualizations. Really showcases how Kedro's modular approach works well for bioinformatics pipelines that involve external API calls and long-running analyses. Would love to hear if others in the community are using Kedro for similar scientific workflows! If anyone is interested in the intersection of biology and AI/ML, we'd love to invite you to our own Slack community at https://biolm.ai/community/ - it's a great space for sharing projects, getting help, and staying updated on bio AI developments!

https://materializedview.io/p/mcp-server-could-have-been-json-file 🤔 Edit: Oops, this should have been in #C03QF15L1K9, but I can't delete it now. facepalming