Staff ML Engineer — Compliance Automation (Remote)

Neobankbeat; United States; À distance; USD 218 000 - 257 000; Plein temps

Avantages offerts par ce poste

Remote-friendly environment

Quarterly in-person surges

Résumé du poste

Coinbase is seeking a Staff Machine Learning Engineer to lead the Compliance Automation team within the Platform group. You will architect production Agentic AI systems, own ML direction for compliance agents, and collaborate with Compliance, Legal, and Operations to automate regulatory workflows at scale.

The role emphasizes auditable outputs, guardrails, and measurable improvements in workflow efficiency, accuracy, and governance. Remote-first with in-person surges, United States.

Qualifications

• 8+ years in machine learning and software engineering, shipping LLM-based or Agentic AI to production.
• Deep expertise in LLM development including prompt engineering, grounding, hallucination reduction, and agent orchestration.
• Experience building evaluation/measurement systems for ML models in production, with drift and accuracy monitoring.

Responsabilités

• Architect and build production Agentic AI systems for compliance automation.
• Own ML direction for compliance agents, incl. model selection and guardrail design.
• Design and implement evaluation frameworks to quantify performance and auditability.
• Partner with Compliance and Legal to translate workflows into automatable specs.
• Mentor engineers on ML system design and responsible AI practices.

Connaissances

LLM development

Agentic AI

Prompt engineering

RAG

Model evaluation

Python

GenAI frameworks

Regulatory tech

Outils

LangChain

Vector databases

Airflow

Spark

Description du poste

Coinbase is seeking a Staff Machine Learning Engineer to lead the Compliance Automation team within the Platform group. You will architect production Agentic AI systems, own ML direction for compliance agents, and collaborate with Compliance, Legal, and Operations to automate regulatory workflows at scale.

The role emphasizes auditable outputs, guardrails, and measurable improvements in workflow efficiency, accuracy, and governance. Remote-first with in-person surges, United States.