Staff Applied Scientist - Knowledge Graphs & AI

About Outreach Outreach, founded in 2014, is the only complete agentic AI platform for revenue teams. Outreach infuses agentic AI, conversation intelligence, and assistive AI to power hundreds of use cases across revenue motions. From new logo prospecting to expansions, deal acceleration, driving retention, and forecasting, Outreach AI automates workflows and frees sellers to focus on more strategic conversations and actions. Revenue leaders benefit from connected account visibility, performance insights, and higher forecasting accuracy across every GTM team. World leading enterprise organizations use Outreach to power their revenue teams, including Databricks, SAP, Siemens, and Verizon to name a few.

About The Job

• We are looking for an Applied Scientist to join a dynamic and innovative AI platform team that is pushing the boundaries of what’s possible in sales execution. If you are passionate about applying cutting-edge research in knowledge graphs and reasoning systems to real-world problems at scale, this is an exceptional opportunity to shape a core piece of Outreach’s AI architecture from the ground up.
• Our team is building a per-tenant contextual knowledge graph that captures the full complexity of each customer’s sales environment: accounts, deals, contacts, rep behaviors, competitive landscape, and the signals buried in calls, emails, and CRM activity. This graph powers contextual reasoning across the platform, driving next-best-action recommendations, deal risk signals, coaching suggestions, and competitive intelligence. In this pivotal role, you will design the underlying representations, extraction pipelines, and reasoning layers that make this possible, working closely with cross-functional engineering and product teams to deliver innovative, scalable, and reliable AI capabilities with direct impact on revenue outcomes.

Your Daily Adventures Will Include Key Responsibilities:

• Knowledge Graph Design & Construction: Architect and evolve per-tenant knowledge graph schemas, including entity resolution, temporal modeling, and ontology design tailored to sales execution domains.
• Information Extraction: Architect NLP pipelines that extract structured knowledge from unstructured conversational and document data (sales calls, emails, CRM notes), including coreference resolution, relation extraction, and event detection.
• Contextual Reasoning & Recommendation: Design reasoning and inference layers over the knowledge graph to power next-best-action suggestions, deal risk scoring, coaching recommendations, and competitive intelligence surfaces.
• Representation Learning: Design and train graph-based models (GNNs, relational embeddings, link prediction) over heterogeneous, multi-relational graph structures to support downstream reasoning and retrieval tasks. Diagnose and address embedding quality issues including cold-start entities, and temporal drift.
• Domain Modeling: Formalize sales execution concepts such as deal stages, buyer engagement patterns, rep behaviors, and account health, into structured representations that ground the platform’s AI capabilities. Extract ontology structure. Lead ontology versioning and migration.
• Cross-functional Collaboration: Partner with engineering, product, and data teams to bring models from prototype to production, ensuring reliability and measurable impact at scale.

Qualifications Our Vision of You:

• PhD in a relevant field such as Computer Science, NLP, Machine Learning, or a related discipline with a focus on knowledge representation and reasoning, information extraction and relationship extraction, graph neural networks, recommendation systems, or conversation AI and dialogue systems.
• Strong engineering fundamentals. You can write production-quality code, not just prototype notebooks. Proficiency in Python; and graph databases or query languages (e.g., Neo4j, SPARQL, Cypher) is required.
• Comfort with ambiguity. You can take a vague product goal and decompose it into concrete technical problems. You don’t need a fully scoped spec to start making progress.
• A track record of building things: whether that’s research prototypes that went beyond the paper, open-source contributions, or side projects that required real systems thinking. You understand the gap between a research prototype and a reliable production system, such as monitoring, data drift, latency, and operational excellence.
• Strong Ownership: Take end-to-en…