The schedule is live. Now the question is: what will your three days at ODSC AI West look like? With 200+ hands-on sessions across 9 tracks, you can build an agenda around the work you actually do and the problems you need to solve next. Explore sessions on agentic AI, AI engineering, evaluation, ML, data infrastructure, Physical AI, developer tools, AI leadership, and more. Whether you’re building models, deploying agents, engineering AI systems, leading technical teams, or deciding where AI belongs in the business, there’s a track built around the work. Take a look at the schedule and build your agenda: hubs.li/Q04BxBVt0 🔥Save on Door Prices -30% Off Ends Friday!

Media mix modeling can answer a critical question: which channels are actually driving sales? The problem is that rigorous Bayesian MMM is often slow and complex to run. At ODSC AI West 2026, @lfiaschi86, Partner Gen AI Vertical at @pymc_labs, and Carlos Trujillo, Principal Data Scientist at @pymc_labs, will present “MMM-PFN: Foundation Models for Bayesian Media Mix Modeling.” Luca brings 15+ years of experience leading data and AI teams across companies including Mistplay, HelloFresh, Alibaba, and Rocket Internet. Carlos specializes in Bayesian statistics and is a core contributor to PyMC-Marketing, working on media mix modeling, budget optimization, and causal inference. In this talk, they’ll introduce MMM-PFN, a foundation model trained on millions of synthetic marketing scenarios to estimate channel contribution and ROAS in a single forward pass, without fitting a new model for every dataset. Join this Data Science & Machine Learning talk at ODSC AI West 2026: hubs.li/Q04C740s0

Learn how to build an agent eval suite before you ship by testing tool use, verifying outcomes, calibrating graders, and setting release gates for more reliable AI agents. #DataScience #AI #ArtificialIntelligence #ODSC #AIAgents #AIEvaluation #MLOps #AIEngineering hubs.li/Q04C6Qk60

Coding agents can write files, install dependencies, and build containers. The harder question is: what should they be allowed to touch while they work? At ODSC AI West 2026, @RamiKrispinAI, Senior Manager, AI, Data Science & Engineering at @Apple, will lead the workshop “Running AI Agents Safely with Docker Sandboxes.” Rami is a senior data science and engineering leader, Docker Captain, LinkedIn Learning instructor, and author whose work spans MLOps, forecasting, data pipelines, and production AI systems. In this beginner-friendly session, attendees will explore how Docker Sandboxes, containers, and microVM-based environments can help isolate coding agents, control access to files and Docker resources, and verify what an agent can actually do. The workshop also includes a practical build of a Streamlit dashboard using a coding agent. Join Rami at ODSC AI West 2026: hubs.li/Q04C7PZ40

What if AI agents could manage the model lifecycle, not just call a model API? At ODSC AI West 2026, @sagecodes, AI Engineer at @union_ai, will lead the hands-on workshop “Own Your AI: Building an Agentic Model Factory with LangGraph and Flyte.” Sage has taught thousands of developers across Python, machine learning, MLOps, and AI observability. In this session, he’ll show how to combine LangGraph and Flyte to build an “ML Engineer” agent that can evaluate open-source models, fine-tune promising candidates, deploy the best performer, and monitor results for future improvements. Attendees will also explore durable execution, distributed CPU/GPU workloads, infrastructure-aware provisioning, and failure recovery, bringing agentic automation into the ML lifecycle itself. Join Sage at ODSC AI West 2026: hubs.li/Q04C73fn0

Meet @runware on October 27–29 in San Francisco! Runware believes the future of AI is modular, so they rebuilt the stack. Their Sonic Pods are a new breed of modular AI data center, deployed as a distributed network to bring more compute capacity online while cutting the cost in half. @runware gives developers one managed API for all AI, while Runware Serverless lets AI engineers run their own workloads directly on Sonic Pods, fully on demand, without managing hardware or committing to multi-year reserved capacity deals. Explore the ODSC AI West Expo: hubs.li/Q04C00cM0

Discover ten confirmed ODSC AI West 2026 sessions AI engineers shouldn’t miss, from LLM inference and agent harnesses to silent failures, MCP clients, and coding agents. #DataScience #AI #ArtificialIntelligence #ODSC #ODSCWest #AIEngineering #LLMs #AIAgents hubs.li/Q04C6QkH0

Prior-data-fitted networks train once on simulated data, then return a posterior in one pass. Carlos Trujillo will show how MFM brings this to marketing mix modeling, and where it falls short, at @_odsc AI West, Wed 28 Oct, 4:05 PM Reply "MFM" to hear when we open-source it.

Excited to share I'd be talking all about need, fun, challenges around building an evaluation for AI Researcher Agents. Tune in to learn more!

AI can automate tasks. But which tool actually helps your team work smarter? Explore how Snowflake CoCo and Claude Cowork compare when it comes to collaboration, automation, and real-world team workflows. #DataScience #AI #ArtificialIntelligence #ODSC hubs.li/Q04C07VT0

The hardest AI benchmarks may come from the people who already know exactly where models fail. At ODSC AI West 2026, Brian Smith, Director of Solutions for GenAI Projects at @iMeritDigital, will present “Building Hard Benchmarks with Academic Experts: Best Practices from a Phonology Eval.” Brian combines hands-on GenAI work with a PhD in Linguistics and 10+ years of experience analyzing language data and building predictive models. In this talk, he’ll share how iMerit is working with academic experts to create more demanding reasoning benchmarks, using a phonology evaluation as the case study. He’ll cover expert recruitment, benchmark design, theoretical guardrails, model-breaking problem creation, validation, and collaborative workflows that turn domain expertise into useful evaluation data. Join Brian at ODSC AI West 2026: hubs.li/Q04C0vD20

What happens when an AI agent leaves the test environment and meets real users? That’s where the Women in AI sessions at ODSC AI West 2026 get practical. 🔹 Tanvi Kopardekar, Invoca - From Demo to Production: Shipping Agentic AI That Real Users Depend On 🔹 Devi Parikh, Yutori - Building and Evaluating Computer-Use Agents 🔹 Kerstin Frailey, Stealth Startup - Why AI Evals Belong to Leaders 🔹 Vyoma Gajjar, ServiceNow - Evaluating AI Agents in Production 📍San Francisco & Virtual | October 27–29, 2026 🔥Save on Door Prices -30% Off Ends Friday! Explore the program and register: hubs.li/Q04C212S0

AI applications can only reason with the context they receive. If that context is hours old, the model is already working with an outdated version of the world. At ODSC AI West 2026, Sujay Jain, Senior Software Engineer at @netflix, will present “Fresh Context at Scale: Streaming Pipelines That Keep AI Applications Connected to a Changing World.” Sujay has helped build Netflix’s real-time data infrastructure, including its Streaming SQL platform and Data Mesh, which processes around 14 trillion events a day across recommendations, search, live events, games analytics, and advertising. In this talk, he’ll share practical patterns for building real-time context pipelines using Apache Flink, Kafka, CDC, Flink SQL, Iceberg lookup joins, data contracts, autoscaling, and observability, from capturing changes at the source to delivering fresh context where AI applications need it. Join Sujay at ODSC AI West 2026:  hubs.li/Q04C0s6b0

What should you do when the same AI agent gives two different answers to the same question? At ODSC AI West 2026, Diane Lin, Tech Lead at @datadoghq, will present “Why Your Agent Disagrees With Itself (And What To Do About It).” Diane leads work on self-improving AI agents for security teams and brings experience across Datadog, Culminate, Zscaler, Amazon Alexa, and Vicarious. In this session, she’ll show why agent inconsistency often comes from ambiguous decision boundaries rather than simple model failure and how teams can use active learning, semantic memory, episodic memory, and targeted human review to make agents more consistent over time. Join Diane at ODSC AI West 2026: hubs.li/Q04C0qMQ0

AI For Work nitter.siddz.com/i/broadcasts/1yxBePgBl…

How can academic experts help build AI benchmarks that challenge even top models? Brian Smith of iMerit, an EXL company, explores this at @_odsc WEST 2026, sharing lessons from a phonology evaluation. Oct 27 | 11:05–11:35 AM #AI #AIEvaluation #ODSC

Explore how @NetApp’s @davidvonthenen designs working, episodic, and semantic memory for AI agents in shared systems. It’s a preview of his ODSC AI West 2026 workshop on memory lifecycle policies. #DataScience #AI #ArtificialIntelligence #ODSC #ODSCWest #AIAgents #AgenticRAG #AIMemory hubs.li/Q04C07V10