Applied AI Scientist

Who We Are: Prior Labs is building foundation models for spreadsheets and databases - the backbone of science and business. While foundation models have transformed text and images, structured data has remained largely untouched. We're tackling this $100B+ opportunity to revolutionize how we approach scientific discovery, medical research, financial modeling, and business intelligence.

Our Momentum: We're the world-leading organization working on structured data, and we're accelerating fast. Our TabPFN v2 model, recently published in Nature, sets the new state-of-the-art for structured data. We've hit 2.5M+ downloads, 5,000+ GitHub stars, and growth is accelerating. We're now building the next generation of models that combine AI advancements with specialized architectures for structured data.

The Team: We’re a small, research-first team, 20 permanent hires selected from 5,000+ applicants - building the next generation of foundation models for structured data. Led by the founders behind the TabPFN lineage, we bring together talent from Google, Apple, Amazon, Microsoft, G-Research, Jane Street, and CERN, among others. We’re backed by heavyweight advisors including Bernhard Schölkopf and Turing Award winner Yann LeCun. Meet the team here.

What's Next: Backed by top-tier investors and leaders from Hugging Face, DeepMind, and Silo AI, we’re scaling fast. This is the moment to join: help us shape the future of structured data AI. Read our manifesto.

You'll be among the first scientists developing an entirely new class of AI models. Our latest breakthrough (TabPFN) outperforms all existing approaches by orders of magnitude - and we're just getting started. This is a rare opportunity to:

• Work on fundamental breakthroughs in AI, not just incremental improvements

• Shape the future of how organizations worldwide work with their most valuable data

• Join at the perfect time: We just received significant funding, have strong early traction, and are scaling rapidly

We're pushing the boundaries of what's possible with transformer architectures for structured data. Key challenges include:

• Scaling our transformer architectures from 10K to 1M+ samples while maintaining performance

• Building multimodal models that combine text and tabular understanding

• Developing specialized architectures for time series, forecasting, and anomaly detection

• Creating efficient inference methods for production deployment

• Researching causal understanding in foundation models

• Designing novel approaches for handling multiple related tables

• PhD in Computer Science, Applied Mathematics, Statistics, Electrical Engineering, or a related field

• Deep experience with ML frameworks, especially PyTorch and scikit-learn

• Strong engineering fundamentals with excellent Python expertise

• Experience in data-science and working with tabular data or time series

• Publications at top-tier venues (NeurIPS, ICML, ICLR) or significant open-source contributions

• Offices in San Francisco, NYC, Berlin, and Freiburg, with flexibility to work across our locations

• Competitive compensation package with meaningful equity

• 30 days of paid vacation + public holidays

• Comprehensive benefits including healthcare, transportation, and fitness

• Work with state-of-the-art ML architecture, substantial compute resources and with a world-class team

• We believe the best products and teams come from a wide range of perspectives, experiences, and backgrounds. That’s why we welcome applications from people of all identities and walks of life, especially anyone who’s ever felt discouraged by "not checking every box."

• We’re committed to creating a safe, inclusive environment and providing equal opportunities regardless of gender, sexual orientation, origin, disabilities, or any other traits that make you who you are.