Senior Machine Learning Engineer II
POS-31662
Role Summary
HubSpot is building the next generation of AI experiences across our go-to-market products. We're hiring a Senior Machine Learning Engineer II to join the Flywheel Context (Contacts) team, where you'll build the context platform layer that powers accurate, high-performing AI assistants and agents.
This is a backend-leaning ML engineering role focused on shipping production software-designing systems that help other engineering teams retrieve relevant facts, manage long/complex context, and evaluate quality at scale. If you love building durable platforms (not just prototypes), this role is for you.
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.
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).
