--- title: "Tactfully taking teams to triumph" newsletter: "MLOps Community" date: 2024-05-02 source: https://aaif.live/newsletters/mlopscommunity/2024-05-02-tactfully-taking-teams-to-triumph --- # Tactfully taking teams to triumph *Plus, check yourself before you wreck yourself, LLMs are fantastic but search is Elastic, and four fun finds* *MLOps Community — Agentic AI Foundation, 2024-05-02* Hot on the heels of the Humane Ai Pin and Rabbit R1, we’re excited to announce ArcaVacua, our own AI in a box! For just $1,000* you can have your own personal assistant, in an extra little thing to carry around, that can make calls, take pictures and even access the internet! But it’s so much more - ArcaVacua will be able to book flights and hotels, automatically reply to messages, detect if that meeting should really just be an email, file your taxes, do your laundry, and cook your rice perfectly**. Exclusive review by Marques Brownlee here! [https://media.giphy.com/media/v1.Y2lkPTc5MGI3NjExN2N6cnRuYjI4Znhma3FjeTFoZWtsNXN2eXYxYWtoamxkeDNzNDBiZyZlcD12MV9pbnRlcm5hbF9naWZfYnlfaWQmY3Q9Zw/xT5LMAmbtURA17LfH2/giphy.gif] *plus monthly subscription of $70 ** some features may not be available at launch, but we 100% pinky promise to release them in future updates at some unspecified time ## Leading Enterprise Data Teams // Sol Rashidi // MLOps Podcast #227 Sol knows how to tackle things. And I don't just mean because she's a rugby player. 🏉 When she's got an issue she uses a "Complexity versus Criticality" framework to prioritize use cases to know where to direct resources: * Complexity involves the difficulty of the problem, factoring in technical challenges and data issues. * Criticality gauges the importance to the business, considering impact on performance and strategic goals. Focus on the critical but simpler cases for quick wins, and plan your resources for those with higher complexity and criticality so they don’t get stuck in perpetual POC purgatory. Another takeaway was the importance she places on relationships. When someone who’s got Sol’s experience leading data teams, some over 800 big, takes the time to do one-to-ones, you know it’s got to be important. Click below to give the podcast a try! (That's like a rugby touchdown). Video [https://home.mlops.community/home/videos/leading-enterprise-data-teams] || Spotify [https://open.spotify.com/episode/3BqKiEaCrxJz4fIkTusbvg?si=5FphMZNCTcyxPW01wEhx4A] || Apple [https://podcasts.apple.com/us/podcast/leading-enterprise-data-teams-sol-rashidi-227/id1505372978?i=1000653725937] [https://podcasts.apple.com/us/podcast/leading-enterprise-data-teams-sol-rashidi-227/id1505372978?i=1000653725937](https://podcasts.apple.com/us/podcast/leading-enterprise-data-teams-sol-rashidi-227/id1505372978?i=1000653725937) [https://home.mlops.community/home/videos/leading-enterprise-data-teams](https://home.mlops.community/home/videos/leading-enterprise-data-teams) [https://open.spotify.com/episode/3BqKiEaCrxJz4fIkTusbvg?si=5FphMZNCTcyxPW01wEhx4A](https://open.spotify.com/episode/3BqKiEaCrxJz4fIkTusbvg?si=5FphMZNCTcyxPW01wEhx4A) ## Handling Multi-Terabyte LLM Checkpoints // Simon Karasik // MLOps Podcast #228 Check – check – 1 – 2 – 300 billion. That’s not me warming up the mic for our in-person conference (June 25 [https://www.aiqualityconference.com/], FYI). It’s more like Simon telling me about how to deal with checkpoints when training a 300 billion parameter model. He creates checkpoints every 1-2 hours during training. It's a balancing act, as you don’t want to slow down training too much. He explains the optimization process of breaking these checkpoints into smaller parts, saving, and loading them in parallel. He also discusses the optimization of CPU usage, the importance of proper model cleaning, and the differences between pre-training and fine-tuning models in ML operations. Plus, he covers the application of various tools such as Kubernetes, Argo, Prometheus, and Weights and Biases for model training, orchestration, and system monitoring. Check the video to check your checkpoints. Video [https://home.mlops.community/home/videos/handling-multi-terabyte-llm-checkpoints] || Spotify [https://open.spotify.com/episode/4Des55VAgBVxoxYN8d6QeA?si=XxbahHV_RjO5-OjpdVphjA] || Apple [https://podcasts.apple.com/us/podcast/handling-multi-terabyte-llm-checkpoints-simon-karasik-228/id1505372978?i=1000654081137] [https://podcasts.apple.com/us/podcast/handling-multi-terabyte-llm-checkpoints-simon-karasik-228/id1505372978?i=1000654081137](https://podcasts.apple.com/us/podcast/handling-multi-terabyte-llm-checkpoints-simon-karasik-228/id1505372978?i=1000654081137) [https://home.mlops.community/home/videos/handling-multi-terabyte-llm-checkpoints](https://home.mlops.community/home/videos/handling-multi-terabyte-llm-checkpoints) [https://open.spotify.com/episode/4Des55VAgBVxoxYN8d6QeA?si=XxbahHV_RjO5-OjpdVphjA](https://open.spotify.com/episode/4Des55VAgBVxoxYN8d6QeA?si=XxbahHV_RjO5-OjpdVphjA) ## Job of the Week [https://ouryahoo.wd5.myworkdayjobs.com/en-US/careers/job/United-States-of-America---Remote/Sports-AI-backend-Engineer_JR0024215?source=LinkedIn](https://ouryahoo.wd5.myworkdayjobs.com/en-US/careers/job/United-States-of-America---Remote/Sports-AI-backend-Engineer_JR0024215?source=LinkedIn) ## Generative AI & Elastic Observability // Anna Maria Modée // IRL #75 Stockholm ## MLOps Community IRL Meetup Generative AI & Elastic Observability // Anna Maria Modée // IRL #75 Stockholm If you’ve been looking for a talk about enhancing search capabilities, look no further! This chat looks at GenAI applications in managing large data repositories and enhancing search capabilities at Elastic. Some key topics include the implementation of supervised machine learning, vector search, and contributions to open-source projects. They also discuss the applications of GenAI in observability and security to showcase its potential in detection, diagnostics, remediation, and generating regulatory reports. I hope this finally helps Bono. Watch it here [https://home.mlops.community/public/videos/generative-ai-and-elastic-observability] [Watch it here](https://home.mlops.community/public/videos/generative-ai-and-elastic-observability) ## Hidden Gems ## IRL Meetups Montreal [https://www.meetup.com/montreal-mlops-community/events/300252938/] - May 2 Stockholm [https://www.meetup.com/stockholm-mlops-community/events/300530812/] - May 7 (📣 shout out to MongoDB) Copenhagen [https://www.meetup.com/copenhagen-mlops-community/events/300447488/] - May 8 (thanks to Todai 🙌) Bristol [https://lu.ma/93ew7j0o] - May 9 (cheers to BJSS 🍻) Denver [https://www.meetup.com/denver-mlops-community/events/300435484/] - May 14 Scotland [https://www.meetup.com/scotland-mlops-community/events/300633800/] - May 16 Amsterdam [https://www.meetup.com/amsterdam-mlops-community/events/300696760/] - May 28 Munich [https://www.meetup.com/munich-mlops-community/events/300673506/] - May 28 [Montreal](https://www.meetup.com/montreal-mlops-community/events/300252938/) Thanks for reading. See you in Slack [https://go.mlops.community/slack], YouTube [https://www.youtube.com/channel/UCG6qpjVnBTTT8wLGBygANOQ?view_as=subscriber], and podcast [https://home.mlops.community/public/content/] land. Oh yeah, and we are also on X [https://twitter.com/mlopscommunity]. The MLOps Community newsletter is edited by Jessica Rudd [https://www.linkedin.com/in/jmrudd/]. --- Source: https://aaif.live/newsletters/mlopscommunity/2024-05-02-tactfully-taking-teams-to-triumph