Principal Data & Applied Scientist – Ontologies, Knowledge Graphs & AI Research

SAP IT Business Systeme•Palo Alto, CA

6d•$198,200 - $420,000•Hybrid

About The Position

SAP Research & Innovation North America is seeking a principal-level individual contributor to define and lead research on learning business semantics from enterprise data and to work with product teams to put research results into use. This is a research and modeling role, not product analytics. The goal is to help AI agents understand business objects, their relationships, rules, authorizations, and decision history so they can reliably act on business data. The role will set a focused research agenda, evaluate emerging research, and develop methods for learning and evolving business ontologies and knowledge graphs from SAP metadata, master data, documents, and process data. Areas include graph-grounded retrieval, semantic parsing into SQL, SPARQL, and OData, API discovery across large schemas, governed agent actions, graph learning, relational foundation models, and evaluation benchmarks and simulation environments for agents working with ERP data. The scientist will collaborate with data teams, domain experts, product teams, universities, and research institutes; contribute to co-innovation projects; publish research; release code and benchmarks where possible; and mentor scientists and PhD interns. This is a hands-on role that includes writing code and running experiments, while driving standards and research directions adopted by other teams. The position is hybrid, with three days per week in the Palo Alto office, and requires travel of up to 10%.

Requirements

Typically 8 or more years of experience in applied research, data science, or machine learning, or a PhD and 5 or more years of experience, including research leadership that changed a product or platform.
An MS or PhD in computer science, machine learning, statistics, mathematics, or a related field, or equivalent practical experience.
Deep expertise in one listed technical area and hands-on experience in another, with at least one area involving knowledge graphs, ontologies, or semantic data integration.
For knowledge graphs, ontologies, and semantic data integration: experience with technologies or methods such as RDF, OWL, SHACL, SPARQL, property graphs, ontology and schema matching, entity resolution, or semantic layers; experience making modeling decisions such as entity boundaries, inheritance, and relationship semantics.
For machine learning on relational, graph, or tabular data: experience with GNNs, relational deep learning, or relational or tabular foundation models.
For LLM systems over structured data: experience with retrieval, tool and API discovery, text-to-SQL and schema linking, tool-using agents, fine-tuning, or post-training.
For AI evaluation: experience with evaluation and benchmark design for AI systems.
Strong foundations in machine learning and statistics.
Fluency in Python and SQL and experience with PyTorch or JAX.
Ability to work with schemas spanning thousands of related tables.
Strong problem-solving skills, including turning ambiguous business questions into testable research problems, starting with a simple viable approach, and stopping or redirecting work when evidence warrants it.
Strong communication skills, including explaining trade-offs clearly to executives, producing written work that informs decisions outside one's team, and working through designs with skeptical engineers.
Comfort using AI tools, including agentic AI, in day-to-day work, evaluating AI-generated outputs, and driving adoption of AI practices within the team.

Nice To Haves

Publications at venues such as NeurIPS, ICML, ICLR, KDD, ACL/EMNLP, SIGMOD/VLDB, ISWC, or WWW.
Widely used open-source software, datasets, or semantic standards work.
Experience with LLM-assisted ontology learning, neuro-symbolic methods, agent memory, context graphs and decision traces, or reinforcement learning for tool-using agents.
Experience with process mining, including object-centric event data, causal inference, forecasting, privacy-preserving data, or synthetic data.
Familiarity with enterprise operations such as finance, supply chain, procurement, or HR.
Experience mentoring scientists or leading university collaborations.
SAP experience.

Responsibilities

Define and lead a focused research agenda, establish success criteria for research directions, and stop or redirect efforts when evidence does not support them.
Track and test emerging research, including relational and graph foundation models, neuro-symbolic agents, and context graphs, using SAP data and make recommendations on future research bets.
Develop methods for learning, aligning, and evolving ontologies and knowledge graphs of business objects, rules, and actions from SAP metadata, master data, documents, and process data.
Develop approaches for agents grounded in business semantics, including graph-grounded retrieval, semantic parsing into SQL, SPARQL, and OData, API discovery across large schemas, and checks that plans respect business rules and user authorizations.
Research graph learning and relational foundation models to predict, explain, and simulate business outcomes.
Create benchmarks and simulation environments for evaluating the accuracy and reliability of agents operating on ERP data.
Work with data teams and domain experts on co-innovation projects to turn enterprise data problems into ontology problems.
Collaborate with teams behind SAP Knowledge Graph, SAP Business Data Cloud, and Joule to deliver research results in products or pilots.
Publish research, release code and benchmarks where possible, co-lead university collaborations, and represent SAP research externally.
Mentor scientists and PhD interns, and help other scientists improve their work.
Write code and run experiments as part of hands-on research.