Specialty
The role
Coinbase is hiring a Machine Learning Engineer for the CX Intelligence Engineering team, which builds the multi-agent platform behind Coinbase Chat, the Help Center and support agent tooling. The role focuses on LLM-based agents for customer support.
What you will do
- Design multi-agent systems with orchestration frameworks such as LangGraph and Google ADK to automate support procedures
- Build integrations using Model Context Protocol to connect LLMs to internal APIs, databases and third-party tools
- Build LLM-as-a-judge evaluation pipelines to monitor agents in production
- Apply RAG, fine-tuning and prompt engineering so chatbot answers are grounded and policy compliant
- Ship resilient, low-latency Python services
- Work with Conversation Design and Product to turn business logic into agent procedures
What they ask for
- 3+ years building and scaling ML or AI systems in production, including agentic ReAct-style systems
- Understanding of the LLM lifecycle: context window management, token optimization and structured output parsing
- Multi-agent orchestration experience (LangGraph, Google ADK or similar)
- Strong Python with microservices, async programming and high-throughput APIs
- Experience with vector databases such as Pinecone or Weaviate
- Ability to explain model behavior to non-technical stakeholders in CX, Legal and Product
What they offer
- Equity and bonus eligibility on top of base salary
- Medical, dental and vision coverage
This is our own short reading of the posting. The full description, and anything that has changed since, is on Coinbase's careers page behind the apply button.
