Sr Machine Learning Engineer

About the Role

Uber's Core Security Engineering org is re‑imagining how identity is protected and managed by fusing large‑scale machine learning with modern agentic AI. As a Senior Machine Learning Engineer you'll lead two complementary tracks: building context‑aware authentication models that decide when and how to challenge users, and creating an agentic AI workflow engine that automates IAM operations-think an LLM‑powered copilot that files, executes, and verifies tasks such as user provisioning, access reviews, and policy changes. No prior security background is required; what matters is your ability to ship reliable, production‑grade ML systems.

What You'll Do

• Risk‑based authentication - Design and deploy models that fuse behavioral biometrics, device posture, and network context to produce dynamic risk scores and trigger step‑up factors or session revocation in real time.
• Agentic AI workflow engine - Architect an LLM‑driven orchestration platform that can observe IAM events, plan multi‑step remediation or provisioning flows, and take safe actions across systems such as Okta, Active Directory, Duo, and ServiceNow.
• Productionize models and agents in high‑scale, low‑latency environments handling millions of auths and thousands of automated tickets per day.
• Build continuous feedback loops so both models and agents learn from outcomes and human approvals.
• Collaborate closely with IAM, SSO, infra, and detection teams to integrate your solutions end‑to‑end-from real‑time decisioning to post‑incident forensics.
• Lead design reviews, mentor engineers, and evangelize ML/GenAI best practices across the security and infrastructure orgs.

• 5+ years building ML or applied data‑science solutions, hands‑on experience with generative‑AI or agentic systems.
• Expert in Python, with proficiency in Go or Java; deep familiarity with ML/LLM frameworks such as PyTorch, TensorFlow, JAX, or LangChain.
• Demonstrated success running ML services in production at large scale and low latency.
• Strong communication and leadership skills; proven ability to set technical vision and influence roadmaps.

• Experience with reinforcement learning, retrieval‑augmented generation, or structured tool‑use frameworks for AI agents.
• Hands‑on work with risk scoring, anomaly detection, or fraud prevention pipelines.
• Familiarity with identity systems (Okta, Entra ID/AAD, SSO, MFA) or enterprise workflow tools (ServiceNow, Airflow) is a plus-but not required.
• Publications, open‑source contributions, or patents in ML, GenAI, or security automation.