DAGWorks Inc.

Hamilton just keeps getting better, loving this! GS Strats/alumni and beyond: Hamilton's graph node caching is getting very close to lazy evals of SecDB's compute graph - worth checking out! #strat #secdb

Are you in town for #SFTechWeek on Monday? Are you looking for the best place to learn about cutting edge #AI from not 1, not 2, not 3, but 4 different companies? Do you want to see demos from seven (7) of the best up and coming AI tooling companies? Then you won't want to miss out on our SF Awesome AI Dev Tools October event by OSS4AI Come see demos from: - Tim Gilboy - Stefan Krawczyk - Wes Nishio - Uli Barkai - Hemnaa Subburaj - Olaoluwa Ogundeji - Yash Khandelwal As well as talks from: - Shuveb Hussain - Ana Robakidze - John Gilhuly - Sushobhan Ghosh RSVP in the comments - hope to see you all tomorrow!

I recently learned about Hamilton, a Python library to create data transformation DAGs. When your data transformation pipelines are Python-based, this tool seems to really help visualizing dependencies of each transformation. This gives you visibility and clarity in your data flows. Has anybody used it before? If so, how do you like it? ?? Hamilton github link: https://lnkd.in/gHkA8jFq #python #dataengineering #analyticsengineering

I recently learned about Hamilton, a Python library to create data transformation DAGs. When your data transformation pipelines are Python-based, this tool seems to really help visualizing dependencies of each transformation. This gives you visibility and clarity in your data flows. Has anybody used it before? If so, how do you like it? ?? Hamilton github link: https://lnkd.in/gHkA8jFq #python #dataengineering #analyticsengineering

?? TL;DR -- Hamilton now has caching, and it's *really* easy to use! ?? Happy Friday folks! I'm really excited to share out that we have finally released #caching as a first-class component in Hamilton. With sf-hamilton==1.79.0, you can fully solve the problem of wasted recompute and slow iteration times! Thierry Jean has been working on this for a while. It's one of those features that every asset-layer framework should have, but is very hard to get right. While a good experience with caching can save you time, money, and frustration, a bad implementation will inevitably make you distrust the framework you're using. This is why we've made it so it *just works* with a one-line change! While there are a host of customizations available (custom hashing mechanisms, custom behaviors), as well as a series of introspection capabilities (visualize + view logs), the default is extremely simple and should be easy to get started with. This, IMO, is the biggest upgrade to Hamilton since we initially released. I think it's a great reason to switch from custom notebooks/script organization to using #Hamilton. Thanks to our OS community, particularly Gilad Rubin for feedback on the feature/docs, and Michal Siedlaczek for an initial implementation we drew inspiration from. Links in ??!

Big thanks to Stefan Krawczyk and the DAGWorks Inc. team for letting me share some of the exciting work we've been doing at Best Egg (and for their awesome #Hamilton library)! Learn how we are approaching ML feature engineering using Hamilton: https://lnkd.in/d8E54qzE

Super excited for this post from Ryan Whitten from Best Egg on using #Hamilton to build their #feature platform. Even though we're in the #GenAI hypecycle (Hamilton is also used a lot here too!), #ML still matters because it still provides a lot of business value! Read about it here: - https://lnkd.in/d8E54qzE Star Hamilton here: - https://lnkd.in/gixMd-xu

In the space of building #AI applications + #agent workflows, there's a lot of interesting innovation going on. People are wrestling with questions of how to build dynamic workflows -- do you let the LLM decide the shape of your graph? Or do you hardcode transitions/edges/encode dynamism in the actions themselves? ?? As with all questions of design, there's no one-size-fits-all answer, but there are some principles that you can pick up. As we've seen a lot of these recently, we decided to write about them! We're writing a series of posts on design patterns with agentic workflows, and decided to start with something simple but powerful. We'll be using #Burr to illustrate these (complete with sample code), although the lessons should go beyond it. In our first post, we address the question of how to build a reliable tool/function-calling workflow (and what even is the difference between tools, functions, and structured outputs?). Enjoy! We'll be releasing more -- would love any feedback/experiences you've got that could add to the conversation! Are there any design patterns/common challenges you'd be interested in hearing more about? https://lnkd.in/gW9nBZmt

In this episode of Ventures with David, I had the pleasure of speaking with Stefan Krawczyk, a Stanford University School of Engineering alumnus and CEO of DAGWorks Inc., a Y Combinator-backed startup addressing a classic challenge faced by data science teams: orchestrating operations. Stefan created a phenomenal Python integration to help data science teams coordinating joint efforts. We discussed Stefan’s journey from working in industry to becoming an entrepreneur, finding product-market fit, and building tools that deliver real value to tech companies. If you're passionate about tech or curious about data science and entrepreneurship, check out the full conversation on my YouTube channel: https://lnkd.in/gd-qww9X. Stay tuned for more episodes!"

I just launched the AI assistant to Salesteq. Its?#burr-based architecture makes it easy to remold to any customer's needs. Check it out at app.salesteq.ai