Posit PBC

Every year, we open the call for talks and wait to see what the community sends in. Every year, we are surprised. This year's posit::conf(2026) program is the best reflection of that yet. Over 100 accepted talks from practitioners across healthcare, pharma, academia, government, nonprofits, and enterprise, all sharing what they have actually built, broken, and learned along the way. The AI thread runs deep this year, but the angle is refreshing. This community is not here to celebrate the hype. Sessions explore the nuances of AI in data analysis, from where it excels to where human judgment still matters most, how to build reliable evaluation layers for LLM-generated work, and what thoughtful integration actually looks like in practice. If titles like ‘Roll for Wisdom’, ‘Trust but Verify’, and ‘Teach Once Use Often’ are any indication, this is a program built for practitioners who want to think, not just ship. Beyond AI, the program reflects the full breadth of what it means to work in data science right now: Quarto accessibility updates, Positron notebooks, Shiny at scale, SQL as a first-class citizen, small teams outperforming their headcount, and clinical workflows that are genuinely changing patient outcomes. This is a community that builds things that matter. And in September, they are all going to be in the same room in Houston. If that sounds like your kind of September, Super Early Bird pricing ends April 10. It is worth locking in before the rate goes up. Browse the full program: https://lnkd.in/eVAsNz6V Register: pos.it/conf #positconf2026 #rstats #datascience #python #opensource #quarto

What even IS an analytics engineer? On The Test Set, dbt Labs CEO Tristan Handy gives us the real definition. "Somebody who fundamentally enjoys the process of curation as opposed to exploration." While data scientists want to read every page and draw new conclusions, analytics engineers want every book on the shelf in the right order so the next person can find it. It's a whole different obsession. Listen now on Spotify: https://lnkd.in/eEHNbi6p Listen now on Apple Podcasts: https://lnkd.in/eHkwTTjb

🏝️ If you were stranded on a desert island with only one #tidyverse package (and a laptop of course)... which one are you picking? Don't see your favorite? Shout it out in the comments!

We are pleased to announce Quarto 1.9 ✈️ Quarto 1.9 brings powerful new capabilities for publishing, accessibility, and brand management for the data science and research community. • Quarto now supports automated generation of `llms.txt` and `.llms.md` files for websites. This feature is designed to make your documentation and technical content more accessible and readable for large language models (LLMs) and AI-powered tools. • This release introduces the pdf-standard option, enabling PDF/A archival formats and PDF/UA compliance to add screen reader support for figures and cross-referenced equations in both LaTeX and Typst. • Direct deployment to Posit Connect Cloud with the new quarto publish posit-connect-cloud command. • The new “List Tables” syntax allows for the creation of complex table content, such as nested lists or code blocks, using intuitive bullet-point syntax. • Significant improvements have been made to the Typst engine, including professional textbook styling for book projects and more flexible article layouts for margin content and full-width figures. Quarto 1.9 also introduces “Privacy-First” defaults for website cookie consent and the new `quarto use brand` command to keep organizational assets in sync across projects. Learn more about Quarto 1.9 and its full range of capabilities: https://lnkd.in/dKCqHNwt

The question for enterprise CDOs and CAIOs is no longer whether to invest in AI. It's where to focus to maximize outcomes. 🎯 That was the heart of the conversation at CDO Vision LA 2026, where Chetan Thapar, Sr. Product Manager, and Greg Hart, Sr. Partner Development Manager, represented Posit as a proud sponsor and Snowflake partner alongside 30+ senior data and AI leaders last week in Los Angeles. Chetan presented “Data Science, Where your Data Lives” as part of the session "From Productivity Myths to Boardroom Metrics: The Real Levers of AI ROI", making the case that value is captured when AI is: ✅ Embedded in governed data foundations, and grounded in semantic models so AI agents understand the business context ✅ Out-of-the-box, integrated directly into workflows ✅ Tied to explicit economic metrics, not just productivity proxies Chetan demonstrated how Posit + Snowflake delivers on all three with end-to-end AI-driven data science running natively inside Snowflake's security perimeter - featuring a real customer story of governed analytics with reproducible statistical methods A theme that rang true across sessions: the most successful enterprises are moving away from building AI infrastructure from scratch and toward partnering with proven platforms that deliver faster time to value. That is exactly what Posit and Snowflake are built for. One idea that sparked the most conversation during the roundtable was calibrated autonomy. The notion that AI autonomy should be tuned to the task, with high autonomy for well-defined workflows and human-steered collaboration for exploratory analysis, was the counterpoint that resonated most in a room full of leaders navigating the agentic AI wave. Another idea that sparked conversation: deterministic skills — the principle that AI agents should orchestrate analytical workflows, but the statistical methods underneath must be validated, auditable, and reproducible. AI speed with full statistical rigor. A highlight of the day was when Chetan presented the Vision Award to Norma Grubb, CIO and Chief AI Officer at Los Angeles Department of Water and Power. Norma's work advancing responsible AI adoption across critical public sector infrastructure is a perfect example of what it looks like to tie AI to real-world impact at scale. 👏 The energy in that room was candid, practical, and boardroom-ready. Exactly why we show up to events like this. Big thanks to AIM Media House, event host Remy Thellier and Snowflake for a fantastic event. 🙌 Snowflake Partner Network

Announcing orbital 0.5.0: Enhanced Model Deployment for Tidymodels 🛰️ We are over the moon to announce the release of orbital 0.5.0, a significant update for running large-scale ensemble models directly in your database! orbital allows you to translate fitted tidymodels workflows into native SQL expressions - so, model predictions execute directly where your data lives. You can deploy high-performance models in restricted or high-scale environments. • The 0.5.0 release adds support for popular boosted tree engines (`lightgbm`, `catboost`), decision trees (`rpart`), and expands prediction types for existing models (`mars`, `multinom_reg`, `rand_forest`). • It also significantly reduces computational overhead and increases execution speed within the database by using nested `case_when()` statements that eliminate redundant checks. • A new `separate_trees` argument enables database query parallelization for ensemble models, so you can now run complex models that previously hit expression depth limits. • orbital now supports several complex non-linear transformations spline steps from the recipes package. Learn more about orbital 0.5.0: https://lnkd.in/dQaXJVuZ

We recorded an enablement session on the new Positron Notebook Editor, demonstrating how to open your existing .ipynb files directly in Positron, with absolutely no conversion needed, and immediately start working in Python or R. Whether you are a data scientist living in Jupyter every day, or a team leader looking for a more predictable tooling landscape, bring your questions to our LIVE Q&A session tomorrow, April 1st at 11:30am ET. Register here: https://lnkd.in/eSZVNRxK See what is possible when your notebook lives inside a data science IDE!

We’re hiring for summer 2026! Posit is offering four exciting internship opportunities for students and early-career professionals passionate about data science and open source software. Our 10-12 week summer internships (starting May 26) span multiple teams: • Skills and Evals Intern (PyData Team) - Develop AI agent capabilities for Python projects like Plotnine and Great Tables • R Modeling Intern (Tidymodels Team) - Contribute to the tidymodels ecosystem with content development, deep learning models, and survival analysis • Shiny Accessibility and Testing Intern - Audit and improve WCAG compliance for Shiny components • Software Engineering Intern (Posit Connect Team) - Build applications for the Connect Gallery using Python, R, and Quarto We’re proud that many previous interns have become full-time Posit employees, including Thomas Lin Pedersen, Kara Woo, and Simon P. Couch 🩷 Learn more in the blog post: https://lnkd.in/dFjnt9Yv

Where is AI making decisions for you right now? AI compounds, automates, and scales bias. It will create structural, symbolic, and ambient failures we haven’t imagined yet. We must acknowledge, and prepare, for this new reality. Laura Gast reflects on the inevitability and value of failure in both human and AI contexts in her posit::conf(2025) talk, Failure (and Mistakes). Watch the full video here: https://lnkd.in/dkAMz4Zi #AI #Failure #Data #DataScience

Ten days. That's how long you have to lock in the lowest price of the year for posit::conf(2026). Our Super Early Bird rate expires on April 10, and we don't want you to miss it. Here's a snapshot of what's waiting for you in Houston, TX this September 14-16: 🎤 A keynote lineup that actually delivers. We're talking about Wes McKinney (creator of pandas and Apache Arrow) on using AI to improve software quality. Christine Zhang from The New York Times on data storytelling at scale. Emily Riederer from Capital One on production-grade data science. And Sara Altman & Simon Couch from the Posit AI Core team on what thoughtful AI integration actually looks like in practice. These aren't just faces on a stage - at posit::conf, you'll find them in the hallway, at the coffee station, and in the lounge all week. 🛠️ Workshops that change how you work. One full day of hands-on learning with instructors like Hadley Wickham, Jenny Bryan, Max Kuhn, and Mine Çetinkaya-Rundel. Whether you're leveling up your Quarto game, building LLM-powered workflows, or finally going deep on tidymodels, there's a room for you. Can't make it to Houston? A virtual workshop option is available too. 💻 Not joining us in person? A dedicated virtual day on September 14 includes exclusive remote-only talks and a Posit-led workshop, plus livestreams of all keynotes and main sessions on the 15th and 16th. Super Early Bird pricing ends April 10!! If you've been sitting on the fence, this is the deadline that actually matters. Register now before the rate goes up. 👉 pos.it/conf #positconf2026 #rstats #pydata #datascience #python #opensource