PriorLabs

Excited to announce a new Job Opportunity at Prior Labs: we're hiring a Staff Backend Engineer (Founding Team) in Freiburg/Berlin Job description below; please share widely. Know someone perfect for this? We offer a €2,000 referral bonus if we hire your recommendation! AI excels at text/images but so far faltered on structured data. Prior Labs is building truly native Foundation Models to solve this (tech published in Nature, 1M+ downloads). Backed by €9M from Balderton Capital, XTX Ventures & AI pioneers (founders/leaders from Hugging Face, Google DeepMind, Snyk, Black Forest Labs...). We’re seeking a top-tier backend architect & technical leader (5+ yrs) for our founding team. You will own and build our core infrastructure from the ground up, tackling complex challenges in reliably serving novel AI models at scale. This involves architecting robust APIs, productionizing research, mastering cloud infrastructure (AWS/GCP), and defining MLOps best practices. This role demands deep system design expertise and a drive to solve hard problems. Python/FastAPI & ML infrastructure experience highly valued. Define foundational technology, make a massive impact alongside world-class researchers, and earn significant founding equity. Application link in the comments! #BackendEngineering #StaffEngineer #AI #MachineLearning #MLOps #SystemDesign #Python #FastAPI #CloudComputing #AWS #GCP #StructuredData #FoundationModels #Startup #Hiring #BerlinJobs #FreiburgJobs #PriorLabs

We're hiring for a founding product role at Prior Labs! This role is for you if you're technical and have built products for data engineers / scientists, ML engineers or data analysts. As our first hire in product, you'll get the chance to shape the future of foundation models for structured data and change the way we interact with databases and spreadsheets today. Backed by €9M funding (Balderton, XTX Ventures) and co-founders of Hugging Face, Silo AI, Snyk, Tessl, Qonto and DataRobot. 📍 Location: Berlin or Freiburg Join us! https://lnkd.in/eKyjuVK2 #hiring #productmanagement #AI #datascience #foundationmodels #tabpfn

Thank you to all those that attended our event last night! It was a pleasure to introduce Prior Labs, our foundation model, TabPFN and showcase some our use-cases to the broader community. Thanks to our panelists and moderator James Wise for an engaging discussion on the future of the data science! You could feel the energy in the room and we couldn't be more excited about what the future holds 🚀 #TabPFN #DataScience #Machinelearning

Taktile is teaming up with Prior Labs to bring TabPFN—a cutting-edge tabular LLM—to risk, fraud, and compliance teams 🚀🚀🚀 Why it matters: ⚡ Predict missing values instantly—think ChatGPT for tabular data. ⚡ Detect fraud and credit risk anomalies, even in small-data regimes. ⚡ Scale faster with leaner teams—no deep data science expertise required. 👉 Maximilian Eber wrote about how risk teams can put TabPFN to work, check it out! Link to the article is in the comments section.

We've got a great line up for the panel at our London event tomorrow - you don't want to miss this. We'll have: Christopher Lynch, Founder & CEO of AtScale and former chairman of DataRobot Hans Buehler, Co-CEO of XTX Markets and former Co-Lead of Data Analytics at J.P. Morgan Global Markets Frank Hutter, Co-Founder & CEO of Prior Labs Panos Athanasiou, EMEA Head of Innovation at Databricks Register here: https://lu.ma/d64ql04k

We're excited to partner with Taktile to empower risk, fraud, and compliance teams with state-of-the-art tabular LLMs! 🎉 Our model, #TabPFN, is now available on Taktile, bringing unprecedented accuracy and efficiency to tabular data predictions in cases of fraud detection and credit risk anomaly detection. Why should you be excited? ✅ Fast and Accurate Predictions: TabPFN effortlessly predicts missing values, operating much like "ChatGPT for tabular data. ✅ Optimized for Small Data: Excelling in scenarios with limited data, TabPFN significantly enhances fraud detection and credit risk anomaly detection. ✅ Empowering Teams: TabPFN paves the way for smaller and more impactful data science teams, making your workflows up to 90% faster. “I couldn’t be more excited for teams using Taktile to start building with this model and see tangible results from integrating it into their workflows,” explains Maximilian Eber, Co-Founder and CPTO at Taktile. Learn how TabPFN is transforming workflows for risk practitioners, enabling faster, smarter, and more accurate decisions: 👉 https://lnkd.in/ebubFJQj #LLMs #datascience #risk #fraud #compliance #fintech

Read how BostonGene is using #TabPFN to improve patient outcomes and accelerate drug discovery! By integrating TabPFN into its advanced ML platform, BostonGene successfully overcame significant challenges posed by limited data availability and complex, high-dimensional immune cell profiling datasets. BostonGene is a pioneer in precision medicine that partners with institutions such as Takeda, BeiGene and MD Anderson Cancer Center. ✔️ Developed a highly accurate TabPFN classifier (90% accuracy) capable of distinguishing cancer patients from healthy individuals using only peripheral blood immune profiles—without needing tumor data. ✔️ Achieved 90% faster ML workflows by eliminating manual hyperparameter tuning. ✔️ Enabled oncologists to more accurately personalize immunotherapy regimens, significantly enhancing patient outcomes. Michael Goldberg, PhD, VP R&D at BostonGene, stated: “Accurately classifying cancer patients based on peripheral immune cell distribution was remarkably challenging—TabPFN made it a reality.” Arseniy Sokolov, former Data Scientist at BostonGene, further highlighted: “TabPFN eliminated tedious hyperparameter tuning, enabling us to focus on core scientific questions rather than technical optimization. Our model development was accelerated dramatically while maintaining outstanding classification performance.” Explore the full story and learn how TabPFN is redefining precision medicine. Thanks to the BostonGene team for their contributions! #PrecisionMedicine #ImmunoOncology #MachineLearning #CancerResearch #ClinicalInnovation #DrugDiscovery #BiomarkerDiscovery

With just one week to go, the excitement is building up! We're thrilled to remind you that our launch event in London is almost here. Next week, on the evening of 11 March 2025, join us for an exclusive panel discussion on how foundation models for tabular data are transforming industries—from finance and healthcare and beyond. You'll be joined by our attendees from leading companies including Meta, Google DeepMind, Databricks, McKinsey and other AI pioneers including fund managers & tech companies. Expect in-depth insights, real-world use cases where TabPFN is making a difference, and engaging conversations 😊 If you haven’t registered yet, now’s the time to secure your spot for the evening. 📅 Date: 11 March 2025 | London, UK 🔗 Register here: https://lu.ma/d64ql04k Looking forward to seeing you there! #LaunchEvent #TabPFN #DataScience #AI #Innovation

We have a seamless integration of TabPFN with Google Sheets thanks to our community member Benjamin Lalanne! Learn more and get started using the GitHub repo for the TabPFN Google Cloud Function here: https://lnkd.in/eSJiKyxd The video below shows how transactions are instantly classified into merchant categories - no need to leave your familiar Google Sheets interface. What other integrations can you build with the TabPFN API and Google Sheets? 👀 Just some ideas from us: 🦾 tagging support tickets 🧑💻 predicting customer churn ⚡ estimate sales or inventories ... and so much more! Get started with the API here and tell us what you're building with TabPFN in the comments: https://lnkd.in/ecRuinr6 #DataScience #AI #GoogleSheets #BusinessIntelligence #TabPFN

Tabular data often has messy text features — preprocessing them is tedious and time-consuming. Our new #TabPFN 2+ removes that hassle: just feed in your data, and get fast, accurate predictions. No feature engineering required. We’d love your feedback! 📊Notebook comparing TabPFN 2+ with CatBoost, CARTE & more on a text-rich tabular dataset: https://lnkd.in/eGXd6cUb TabPFN 2+ uses semantic representations of text rather than simple target encoding. This lets it generalize to unseen textual features — categories not present in the training set. With this, TabPFN 2+ is similarly strong on textual features unseen during training as on seen ones. Text features are exclusively available via API: https://lnkd.in/ex_vyU52 This also includes free GPU compute (TabPFN is much faster on GPU). 🚀 We're making our models multimodal and are actively hiring - join us on this journey! https://lnkd.in/eFiPgCcg Join our Discord community to discuss anything related to tabular data science and TabPFN: https://lnkd.in/e_tEGNpe