May 2024 DVC Pulse!

???? Hi friend!

And just like that, we’ve made it through May! Here’s what came up in the last month.

?The Latest

?? New Videos

• Achieving Production-level Performance in RAG with DSPy, Parea and DVC - Joschka Braun joined us at this month’s meetup to provide this tutorial on how to integrate these tools for a multi-hop use case in RAG.
• Dmitry Petrov presented Data Versioning in Generative AI: A Pathway to Cost-effective ML at the OSS4AI event at GitHub in San Francisco on May 20th. You can find his presentation here at 1:02 which shows the application of our latest tool to LLMs.

?? What we're looking at

• OpenAI Media Manager - Open AI announced it’s working on it’s Media Manager to give creators and content owners the ability to specify if or how they want their works included in machine learning research and training. 2025 is the target date for the availability of the tool. The research they do and the technology they build should ultimately have implications for personal privacy as well. It could not come soon enough IMO.
• Model Face-off - Open AI released GPT-4o and Google launched Gemini 1.5 Flash+. Thomas Schulz?quickly created?an AI Builders Hackathon to enable builders to face-off the two models to determine which technology is better.

• Founder Docs?- I know many of you have become builders yourselves and are using LLMs to solve some unique problems and start your own companies. Much like that mentioned in the previous bullet point, AI founder incubators and meetups are popping up the world over. If you’ve never been a founder before and were wondering, “Where are the docs for this?” Wing Venture Capital has you covered with their Founder Docs.
• Unstructured Data Solutions - Finding it challenging to organize your pipeline of unstructured data? Jonathan Vickery , Research Associate at Activant sheds some light on challenges and changes to come in the industry to deal with unstructured data in his article,?Unstructured Data Pipelines - Capitalizing on the “junk” of the data world. We’re looking forward to adding our solution to the mix soon!

• Latest from AI Alliance -?Understanding AI Trust and Safety: A Living Guide - The AI Alliance has put out this doc to help you understand potential trust and safety issues, along with their mitigating strategies, and put into place best practices in your AI development. The guide is also full of additional resources along these lines. Join the AI Alliance here.
• Learnings from a Year of Building with LLMs - Hot off the press, Eugene Yan , Bryan Bischof , Charles Frye , Hamel H. , Jason Liu , and Shreya Shankar generously share What we Learned from a Year of Building with LLMs (Part 1).? It is full of great learnings and practices as we all figure out how to move from POCs to production-ready LLMs. This first part focuses on tactics.? Don't miss this one and upcoming parts two and three on the operational and strategic aspects, respectively!

??What's Coming!

???Mark June 25th on the calendar! We will have some good news! And a meetup about the good news on the 26th!

?? Community-Generated Content

? Videos:

• Datamaran on using Argilla and DVC in MLOPs workflows for ESG Governance - Vincent Rizzo and Martín Quesada Zaragoza provide this great overview and tutorial on how they use DVC, CML, and Argilla to help them with their ESG (environmental, social, and governance) use case at Datamaran.
• Build Real-World Machine Learning Project: Step-by-Step Guide using FastAPI, DVC & Poetry - Venelin Valkov provides a great and thorough tutorial. Venelin has a bootcamp and Newsletter as well.
• 25 Best VS Code Extensions for AI (ML, DL, GenAI) Development in 2024 - Sunny Savita provides this great resource video of best VS Code extensions in 2024 - of course DVC made the list. Get ready to supercharge your IDE!
• MLOps Machine Learning na Prática - (Portuguese language) Leonardo Mauro P. Moraes of Amaris Consulting provides an in-depth overview of MLOps, including how data science works, how MLOps supports Data Science, and how to get on the MLOps pathway.
• Data Version Control - DVC - (Russian Language!) this video by Gorbunov Vladislav will give you a great overview of the DVC ecosystem and show you basic workflows. Spacibo Gorbunov!
• Introduction to MLOps Engineering and Real Client Projects - Alan Qian from WeCloudData Academy created this video that teaches about DVC as tool used by experienced MLEs in this intro to MLOps presentation.
• Applied Data Analysis, Lecture 13 - Introduction to MLOps - (Russian Language) - Vladislav Goncharenko - Provides this MLOps overview including DVC, Kubernetes, Lambda, SLURM, MLflow, Airflow, and others.
• Do you use DVC in your ML projects ???- our first Community-produced YouTube short! Thanks for the shoutout Shubham Sharma!

? Articles:

• What is MLOPs and How to Leverage It for Your Machine Learning Projects - Natalia Palamarchuk of Flyaps provides a great overview of MLOps, and provides industry use cases and how they benefit from MLOps.
• MLOps Roadmap [2024]: A Complete MLOps Career Guide - Mayank Gupta offers this nice MLOps Career Roadmap to help you on your way!
• Why are there so many Ops acronyms? - Venkatesh Subramanian writes this nice article about what all these ops monikers mean, what the facets of ops are as it pertains to ML, and what's important.

Thanks for the read!? We'll see you next month!?

To your continued success,

Jeny De Figueiredo ( Jenifer De Figueiredo )

Community Manager, DVC.ai