O'Reilly

A semantic layer can bring a single source of truth accessible to everyone who needs it at the modern enterprise. The promise is compelling. But what happens when organizations actually build and deploy one? To find out, Jeremy Arendt interviewed several early adopters who’ve moved semantic layers from concept to production. Four themes emerged from those conversations: some surprising, some predictable, and a few that will sound familiar to anyone who’s ever shipped data infrastructure. Read on #Radar to learn more: https://lnkd.in/dZVcJbQT

Four new keynote speakers are taking the stage at Ai4 2026, bringing deep expertise across autonomous systems, enterprise-scale AI, world models, and the architecture powering the next era of intelligent machines. Dmitri Dolgov – Co-CEO of Waymo and a founder of the Google self-driving car project Sebastian Thrun – Founding CEO of Waymo & Udacity Jeetu Patel – President & Chief Product Officer at Cisco Tim O'Reilly – Founder & CEO of O'Reilly They join an already stellar lineup including Geoffrey Hinton, Fei-Fei Li, and Andrew Ng. Don’t miss the event where AI decisions are made and real AI use cases are in action. Register now for $1500 off final pricing: https://lnkd.in/gJwZ456b

On March 5th, Rob Reid is joining O'Reilly’s Software Architecture Superstream: Software Architecture and the Age of Agentic AI. This session gets into the technical realities of building for autonomous AI agents, including: -> Architectural tradeoffs when agents operate independently and continuously -> Turning static data into contextual, traceable, and actionable inputs for automation -> Using event-driven patterns and standardized protocols to connect AI workflows with real-world systems Save your spot: https://lnkd.in/eeK7fjTW

The best part of Monster Scale Summit is that no matter where you are, you can learn from over 60 speakers who are solving engineering challenges from a variety of angles. Another benefit is that registrants receive 30-day full access to the complete O'Reilly library and a 45% discount on all Manning Publications Co. books. Here are 11 reads that developers, engineers, and architects can't afford to miss > https://lnkd.in/eB4-qNMJ #ScyllaDB

When #AI makes people more productive, who benefits? Mostly big tech according to O'Reilly. We need to turn this into a revolution that benefits everyone, like the internet to avoid creating a loop of job cuts → less spending → less innovation. →

When we started out three years ago, I never could have guessed this would happen. I'm thrilled to announce that Python Polars: The Definitive Guide has officially been translated into Japanese! Seeing the original English edition sitting right next to its new sibling feels surreal. A massive thank you to the translator, Hidemoto Nakada, for his hard work and for even taking time to correct a few mistakes we made along the way. And special thanks also goes to O'Reilly's editor for this edition, Ryoko Akaike, for sending the complimentary copy Jeroen Janssens and I can use to start learning Japanese ourselves ;) Arigato gozaimasu!

Do you want to learn how AI tools like Copilot and ChatGPT can help you get more out of Excel? If so, I'm running my Generative AI for Excel course for O'Reilly on March 4. Find out more and register: https://lnkd.in/e_RR728Z

O'Reilly courses are always top-notch. MLOps and LLMOps is where workflow and provenance touchpoints can be placed, so the output can later be traced. Very valuable in developing AI systems that meet your guidelines for responsible AI. AI summary of agenda: Day 1: MLOps Foundations (build the basics) What MLOps/LLMOps is and why it exists: ML lifecycle + why production is hard; where LLMs add extra operational complexity. Git for ML teams: managing code + notebooks + model artifacts; branching/commits in practice. Experiment tracking with MLflow: log params/metrics for reproducibility; compare runs. Docker fundamentals: create consistent dev/prod environments; containerize a simple inference API (e.g., Flask + model). End-to-end pipeline design: data ingestion → training → validation → deploy → monitor; simple pipeline demo. CI/CD for ML: automated testing, retraining triggers, and deploying via GitHub Actions. Wrap-up/Q&A. Day 2: Applied MLOps (run it in the real world) LLMOps spotlight (part I): putting LLMs into pipelines; checkpoints + GPUs. Kubernetes for ML: deploy containerized services, scale, and run locally in a cluster. Monitoring & logging: drift/performance/alerts; LLM metrics + feedback loops. Scaling ML systems: throughput, multimodel serving, and resource management (especially GPUs/TPUs). Advanced K8s: high availability, autoscaling, quotas; load testing + HPA demo. Workflow orchestration: Kubeflow Pipelines or Airflow; LLM steps like prompt engineering or fine-tuning. Advanced experimentation: A/B tests, canary releases, bandits; run two model versions and compare performance. LLMOps spotlight (part II): memory footprint, GPU scheduling, user feedback; discussion on how LLM deployment differs. Lifecycle & governance: registries, provenance, retraining triggers, compliance, multi-version + checkpoint management. End-to-end integration demo: update model with new data → rebuild Docker → deploy to K8s; small “LLM update” scenario. Wrap-up/Q&A.

𝐂𝐨𝐦𝐞 𝐭𝐨 𝐦𝐲 O'Reilly 𝐜𝐨𝐮𝐫𝐬𝐞 𝘋𝘢𝘵𝘢 𝘌𝘯𝘨𝘪𝘯𝘦𝘦𝘳𝘪𝘯𝘨 𝘍𝘶𝘯𝘥𝘢𝘮𝘦𝘯𝘵𝘢𝘭𝘴 𝘪𝘯 2 𝘸𝘦𝘦𝘬𝘴 𝐭𝐨 𝐥𝐞𝐚𝐫𝐧 𝐭𝐡𝐞 𝐩𝐫𝐢𝐧𝐜𝐢𝐩𝐥𝐞𝐬 𝐨𝐟 𝐦𝐚𝐤𝐢𝐧𝐠 𝐝𝐚𝐭𝐚 𝐮𝐬𝐚𝐛𝐥𝐞 𝐚𝐧𝐝 𝐚𝐯𝐚𝐢𝐥𝐚𝐛𝐥𝐞 𝐭𝐨 𝐲𝐨𝐮𝐫 𝐨𝐫𝐠𝐚𝐧𝐢𝐳𝐚𝐭𝐢𝐨𝐧: • Discern and organize roles of data engineering 💼 • Manage the data engineering life cycle 📊 • Use the right tool to solve the right type of problem 🛠️ • Do a hands-on exercise with real NOAA tornado data 🌪️ I will also share a black hat story 💻🫆 I was involved in, and do a thought experiment with family photographs to answer the question "What is data?" 📷 https://lnkd.in/gyGmPh3x

Written by 👩🏻💻 Mikiko B. published on O'Reilly, learn why multi-agent systems need memory engineering. Read it here 👉 https://lnkd.in/gKVby4BP #mongodb #nosql #ai #memory #database