Senior AI Researcher – Relational Foundation Models

Description

The team builds foundation models for structured enterprise data. It is launching and continuing to improve its first model, the first release of a relational model line.

Relational foundation models are general-purpose systems that learn, reason and predict from complex data spread across many tables, with no manual feature engineering or task-specific pipelines.

The role covers basic research through to production-ready methods that run on real enterprise databases.

The company builds enterprise applications and business AI across finance, procurement, HR, supply chain and customer experience, and has done so for over 50 years.

You will work with teams around the world, mentor people and take part in global forums. The company's growth culture gives you room to try new methods and support to scale your impact internationally.

Responsibilities

• Do original research on relational foundation models, which are architectures trained on many multi-table enterprise databases and adaptable to different schemas and prediction tasks.
• Design new model architectures, pretraining methods, learning objectives and inference approaches for relational and structured data.
• Run large distributed GPU experiments and build strong baselines, ablation studies and evaluation benchmarks.
• Help build datasets, synthetic data generators and evaluation frameworks for relational learning.
• Turn research results into scalable prototypes that engineering and product teams can integrate.
• Lead research projects and mentor researchers and engineers.
• Publish in top venues such as NeurIPS, ICML, ICLR and KDD, or comparable ones, and engage with the wider research community.
• Keep up with progress in foundation models, tabular learning, graph learning and large-scale machine learning.
• Architect end-to-end solutions that add new product capabilities and operational efficiencies, guide roadmap choices with evidence, lead complex analyses, and make sure models are interpretable, reliable and tied to business outcomes.
• Shape strategy across functions with your findings.

Requirements

• Required: a PhD in machine learning, computer science, statistics, applied mathematics or a closely related field, or equivalent proven research experience.
• Required: a solid publication record at leading venues such as NeurIPS, ICML, ICLR, KDD, AAAI or AISTATS.
• Required: deep knowledge of modern foundation model methods, including Transformers and attention, large-scale pretraining, in-context learning and transfer learning.
• Required: hands-on experience in at least one of tabular machine learning, graph neural networks, relational learning or structured prediction.
• Required: excellent Python and PyTorch skills, with the ability to implement, debug and evaluate research ideas end to end on your own.
• Required: rigorous experimental practice, including strong baselines, controlled ablations, honest failure analysis, and awareness of data leakage and evaluation artifacts.
• Required: the ability to work largely independently, from defining problems through to validated results.
• You design secure, scalable systems, reason about data structures and APIs, and use observability, CI/CD and automated testing to ship safely and quickly. You can tune performance deeply, apply secure-by-default patterns, and mentor others on clean, maintainable code.
• Preferred: experience with relational deep learning, graph neural networks or heterogeneous graph learning.
• Preferred: in-context learning beyond natural language, including meta-learning and amortized Bayesian inference.
• Preferred: creating synthetic data for pretraining and running large-scale pretraining on several datasets.
• Preferred: knowledge of relational database systems, SQL or data warehouse architectures.
• Preferred: distributed training and large-scale GPU experimentation.
• Preferred: benchmark development or contributions to open-source machine learning libraries.

Conditions

The salary range is stated in the job posting. Actual pay falls within that range and depends on education, skills, experience, location and the scope of the role, as decided through selection.

Variable incentives may be offered where applicable, subject to the relevant plan terms. Any incentive is target-based, and actual payouts depend on company and individual performance.

A summary of benefits and eligibility requirements for North America is available.

The company's ethical AI guidelines govern any use of AI in recruitment, and a violation may lead to disqualification from the hiring process.

Only permanent roles are eligible under the employee referral program for employees, and specific conditions may apply to some roles.

The company is an equal employment opportunity employer and provides accommodations to applicants with disabilities on request.

Benefits

• The company describes a culture of inclusion, a focus on health and well-being, and flexible working models.