SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 218,025 - 256,500 / annual
Coinbase is seeking a Staff Machine Learning Engineer to own the identity verification (IDV) ML systems that protect account integrity across millions of signups, account recoveries, and high-risk actions. You will lead the technical strategy for IDV ML end-to-end, from architecture through production enforcement.
Key responsibilities include:
- Own the full IDV ML stack: document authenticity models, 1:1 and 1:N face-matching, liveness detection, presentation-attack detection, and deepfake/injection detection from feature pipeline through threshold tuning and production enforcement.
- Build identity-graph systems using Graph Neural Networks (GNNs) that cluster accounts sharing biometric, device, and document signals to detect synthetic-identity rings and coordinated fraud at onboarding.
- Develop behavioral and device-intelligence models for capture-session anomaly detection, bot-vs-human classification, and device-fingerprint-based risk scoring at real-time latency.
- Drive vendor ML strategy by benchmarking external models against a Coinbase-owned evaluation set, designing dynamic routing logic across providers and geographies, and building the in-house evaluation layer that catches regressions before they reach users.
- Lead and mentor senior and mid-level engineers in the pod while partnering with ML Platform and Risk ML teams to align cross-company ML system design.
Coinbase is a remote-first company with quarterly in-person "surge" sessions for intense collaborative work.
REQUIREMENTS:
- 8+ years deploying production ML systems at scale, with proven technical leadership owning cross-team ML architecture from design through production.
- Domain experience in identity verification, biometrics, or account integrity with deep applied ML in at least two of: computer vision/biometrics, GNNs, sequence models, or NLP/LLMs.
- Expert-level Python with production experience in TensorFlow or PyTorch, including model training, evaluation, and serving infrastructure.
- Track record translating KYC/AML requirements and fraud trends into ML roadmaps and communicating trade-offs to Product, Compliance, Risk, and Security stakeholders.
- Ability to utilize generative AI responsibly, maintaining human oversight to deliver business-ready outputs and drive measurable improvements in workflow efficiency, cost, and quality.
