--- title: "Are you LLM-native?" newsletter: "MLOps Community" date: 2024-10-10 source: https://aaif.live/newsletters/mlopscommunity/2024-10-10-are-you-llm-native --- # Are you LLM-native? *Plus, a tasty podcast for you, a mini summit on optimizing LLMs, a special guest at the next reading group, and hidden gems.* *MLOps Community — Agentic AI Foundation, 2024-10-10* https://go.mlops.community/jxnr44 In our sights this time: AI agents - specifically those in production or just about to be. We're looking beyond experimentation and basic R&D to focus on what it takes to actually deploy buying agents, customer service, analytics tools, or any AI handling multi-step tasks. Join us November 13 for the AI Agents in Production [https://go.mlops.community/jxnr44] virtual conference. Austin Powers fancy dress optional. ## Making Your Company LLM-native // Francisco Ingham // MLOps Podcast #266 “To be native to a place we must learn to speak its language.” The words of Robin Wall Kimmerer talking about respecting the environment, but they fit for this episode. Francisco and I talk about what being "LLM-native" really means. He shares how they’re using LLMs at Pampa Labs, like building small agents for everyday tasks—ordering food, managing team expenses—just to see where they add value. We discuss how to find the balance between where LLMs work well and where humans still do a better job, especially with decision-making and creative work. Francisco’s approach is to test things lightly first, then ramp up evaluation as the solution proves itself. It’s a great listen if you’re into real-world, practical LLM use cases. “You should know the score by now, You're a native New Yorker LLM-er…” Also not my words, the words of Odyssey’s 1977 hit. Video [https://go.mlops.community/7hzlch] || Spotify [https://go.mlops.community/h86hao] || Apple [https://go.mlops.community/3vwd0d] [https://go.mlops.community/3vwd0d](https://go.mlops.community/3vwd0d) [https://go.mlops.community/7hzlch](https://go.mlops.community/7hzlch) [https://go.mlops.community/h86hao](https://go.mlops.community/h86hao) ## Centralized or Decentralized ML Platform? // Jelmer Borst and Daniela Solis // MLOps Podcast #267 To centralize or decentralize ML approaches? They say don’t put all your eggs in one basket, but what if it’s a picnic basket? Daniela and Jelmer from Picnic talked about their journey from a decentralized machine learning setup to a centralized platform. They discussed the benefits of centralization, such as improved collaboration, faster iterations, and consistent model quality across use cases. They also highlighted how clean data, a lightweight feature store, and robust model monitoring tools helped their data scientists scale efficiently while maintaining high performance and innovation. Be smarter than the average bear and click below to listen. Video [https://go.mlops.community/ixh6zs] || Spotify [https://go.mlops.community/dp93ga] || Apple [https://go.mlops.community/lzoqzj] [https://go.mlops.community/lzoqzj](https://go.mlops.community/lzoqzj) [https://go.mlops.community/ixh6zs](https://go.mlops.community/ixh6zs) [https://go.mlops.community/dp93ga](https://go.mlops.community/dp93ga) ## MLOps Community Mini-Summit with SAS Looking for practical strategies to optimize LLM performance? This mini summit covers everything from scaling efficiently to improving structured outputs and reducing token usage—key for deploying LLMs in production. Tom, Advisory Solutions Architect @ SAS, shared how traditional NLP techniques like entity extraction and text analytics can preprocess large datasets, reducing hallucinations and boosting accuracy. Matt, CTO @ Fuzzy Labs, discussed using VLLM for faster inference and RayServe for cost-efficient scaling. Vaibhav, CEO @ Boundary ML, highlighted how BAML generates precise structured outputs, cutting token usage and streamlining pipelines for faster, more reliable data processing. Optimize the next 60 minutes—click here to watch! [https://go.mlops.community/z0orc8] ## Job of the Week [https://go.mlops.community/1dqor2](https://go.mlops.community/1dqor2) ## MLOps Community Reading Group We’re excited to share that our next session on What is the Role of Small Models in the LLM Era: A Survey [https://go.mlops.community/ybxhjl] will feature a special guest—the paper’s author, Lihu! Originally set for today, we’ve moved the session to the October 17 (11 AM ET - 12 PM ET) to accommodate Lihu’s participation. If you’ve already registered, your calendar invite has been automatically updated. This is a great chance to hear directly from Lihu, ask your questions, and explore the findings of the paper. If you haven’t signed up yet, REGISTER HERE [https://go.mlops.community/blepmi] to join us. [What is the Role of Small Models in the LLM Era: A Survey](https://go.mlops.community/ybxhjl) ## Hidden Gems Interested in partnering with us? Get in touch: partners@mlops.community Thanks for reading. See you in Slack [https://go.mlops.community/slack], YouTube [https://go.mlops.community/mgNff7], and podcast [https://go.mlops.community/ttftlf] land. Oh yeah, and we are also on X [https://go.mlops.community/twitter]. The MLOps Community newsletter is edited by Jessica Rudd [https://go.mlops.community/cetjhj]. --- Source: https://aaif.live/newsletters/mlopscommunity/2024-10-10-are-you-llm-native