Senior Machine Learning Engineer II

Required Qualifications

• Strong track record shipping production backend systems as a senior engineer (ownership from design to delivery).
• Professional Java experience building maintainable, testable services in production (this is core to the role).
• Experience implementing ML workflows in production (e.g., retrieval/ranking pipelines, feature/data pipelines, model/embedding services, evaluation frameworks).
• Comfort working with data tooling and data-intensive systems (large datasets, pipelines, and service integrations).
• Experience operating software at meaningful scale (e.g., high throughput, significant data volume, performance and reliability constraints).
• Strong engineering fundamentals: system design, code quality, debugging, observability, and operational excellence.

Nice-to-Have Qualifications

• Experience with search/retrieval/relevance/ranking systems (highly aligned to context work).
• Experience with RAG-style systems, embeddings, vector search, or hybrid retrieval strategies.
• Familiarity with LLM evaluation patterns (golden sets, automated metrics, human review), hallucination mitigation, and quality measurement.
• Experience with distributed systems, event-driven architectures, or stream processing.
• Cloud/platform experience (e.g., Kubernetes, AWS/GCP) and running services in production.
• Some Python experience (useful, but not the primary language).

What You’ll Do

• Design, build, and operate backend services that power context retrieval and enrichment for AI assistants and agents.
• Build platform capabilities for storing, searching, and retrieving “insights” and relevant facts across HubSpot’s GTM data.
• Develop systems to manage and compress context when it gets large (e.g., long contact histories, high-volume CRM data).
• Create tooling that allows other engineering teams to ship assistants/agents faster, with consistent APIs and reusable primitives.
• Build and maintain evaluation and measurement approaches (offline evals, golden datasets, automated metrics, human review loops) to ensure context quality and answer accuracy.
• Collaborate closely with sister platform teams and downstream product engineering teams (your “customers”) to integrate platform capabilities into real experiences.
• Own end-to-end delivery: architecture, implementation, observability, performance, reliability, and iteration.