Prior Labs
Hiring now
Research Scientist, Foundation Model
On-site · Berlin, State of Berlin, GermanyFull-timeUpdated 6 weeks ago
Prior Labs
ApplyPrior Labs
Hiring now
On-site · Berlin, State of Berlin, GermanyFull-timeUpdated 6 weeks ago
Prior Labs
ApplyFoundation models transformed text and images. Structured data - the largest and most consequential data format in the world - stayed untouched, until now. What LLMs did for language, we're doing for tables.
We pioneered tabular foundation models: TabPFN v2 was a Nature cover story, has passed 3.5M+ downloads and 7,500+ GitHub stars, and runs in production from detecting lung disease with Oxford Cancer Analytics to preventing train failures with Hitachi. The hardest problems - millions of rows, real-time inference, entirely new modalities - are still open, and no one else is working on them at this level.
We're a small, highly selective team of 40+ with backgrounds from Google, DeepMind, Meta, Apple, Amazon, Jane Street, and CERN, led by Frank Hutter, Noah Hollmann, and Sauraj Gambhir, and advised by Bernhard Schölkopf and Turing Award winner Yann LeCun.
In July 2026, less than 18 months after our €9M pre-seed, we joined SAP as an independent frontier AI lab - same team, mission, and open-weights models, now backed by more than €1 billion over four years.
We’re hiring researchers to advance foundation models for structured data, with particular interest in structured data, in-context learning, meta-learning, time series and forecasting, relational learning, causal inference, and recommender systems. We welcome deep expertise in any one of these areas, as well as broader foundation-model research experience.
You’ll develop models and learning methods that generalize across datasets and tasks. We’re especially interested in researchers who can bring insights from their field into the design, training, and evaluation of foundation models.
Structured data presents challenges that language and vision architectures don’t solve out of the box. Tables lack a shared vocabulary across datasets; time series introduce temporal dependencies; relational data connects information across tables and entities. TabPFN demonstrated what foundation models can achieve for tabular prediction, and many fundamental research questions remain.
At Prior Labs, Research Scientists drive the core model agenda. You’ll help identify worthwhile research questions, design experiments that distinguish between competing explanations, and take ownership of projects through to publications and model releases. You’ll implement and run your own experiments, interpret the results, and work closely with research engineers to turn promising ideas into reliable models.
The problems you could work on include:
Pushing the limits of dataset size and context length while improving training and inference efficiency.
Building multimodal models that combine tabular data with text and other modalities.
Designing architectures and learning methods for time series, forecasting, and anomaly detection.
Developing relational models that learn across multiple related tables.
Investigating how foundation models can support causal inference and estimation of intervention effects.
Extending foundation-model methods to recommender systems.
Your focus will depend on your strengths and our research priorities. You don’t need expertise across all of these areas or prior experience with TabPFN.
A track record of original research contributions, demonstrated through publications at leading venues in your field, influential open-source work, benchmarks, or deployed methods. We also welcome earlier-career candidates with exceptional academic or independent technical work. We care about your contribution and the quality of the work.
Strong experience developing and analyzing machine learning models, with the ability to implement, train, and debug your own methods in PyTorch.
Solid understanding of training dynamics, generalization, and common failure modes in deep learning systems.
Strong experimental judgment: you choose appropriate baselines, investigate unexpected results, and distinguish robust improvements from evaluation artifacts.
Excellent engineering fundamentals and strong Python skills, with a track record of writing high-quality research code.
Intellectual honesty and constructive collaboration: you explain your reasoning, acknowledge limitations, take feedback seriously, and change your approach when the evidence calls for it.
Research experience in tabular data, time series, relational learning, causal inference, or recommender systems. Depth in one area is valuable.
Experience at an early-stage startup or research lab that regularly releases models or systems.
Contributions to open-source ML libraries or tools.
Experience with model scaling, distillation, inference optimization, or efficient architectures.
You'll work alongside researchers and builders who hold themselves to a very high bar - in the quality of their work and in how they work with each other. We move fast and still take the time to do things right.
Our teams are based in Berlin, Freiburg, and New York - when you're working on something as hard as TabPFN, being in the same room matters. But great people come from everywhere, and in exceptional cases we're open to remote, which usually means frequent travel to one of our offices. Wherever you're based, the whole company comes together regularly for offsites to build and celebrate together.
The best products and teams are built by people with a wide range of perspectives and backgrounds. We welcome applications from all identities and walks of life - especially if you've ever felt discouraged by "not checking every box" - and provide equal opportunities regardless of gender, sexual orientation, origin, disability, or any other trait that makes you who you are.
We care about how your data is handled - see our Recruiting Data Privacy page