ML Engineer

Role Overview

Docker is hiring a ML Engineer as one of the founding engineers for its Intelligence Org (staff-level scope). This is a hands-on build role where you help decide what to build, how to build it, and how it fits into Docker’s broader platform.

Responsibilities

• Design, train, evaluate, and ship ML systems for governance and security capabilities.
  • Start with areas such as:
    • Prompt injection detection
    • Behavioral anomaly detection
    • Trust scoring
    • Policy recommendations
• Build the supporting ML infrastructure, including:
  • Data pipelines
  • Feature stores
  • Model serving
  • Evaluation harnesses
  • Feedback loops to speed up iteration
• Make pragmatic build-vs-buy decisions using frontier models, off-the-shelf tooling, and managed services when appropriate.
• Set technical direction: own architecture, evaluation methodology, model lifecycle, and the bar for shipping.
• Help recruit and mentor as the team grows.
• Participate in a 24/7 on-call rotation for the Agentic Platform and carry pager responsibility (for services you build/operate).

Requirements

• 5+ years of deep applied ML/AI experience with a proven history of shipping production systems.
• Experience in fraud, abuse, safety, security, or trust domains (adversarial dynamics, imbalanced data, high-stakes decisions).
• 4+ years of full-time, hands-on professional software engineering experience (backend/infrastructure/platform).
• Bachelor’s degree in CS/Engineering or equivalent practical experience.
• Built and owned ML-adjacent systems (e.g., data pipelines, serving, evaluation, monitoring) and shipped customer-facing products end to end.
• Strong capability working with production LLM-based systems, including:
  • evaluation, prompt engineering, fine-tuning
  • retrieval and guardrails
  • agent frameworks
• Familiarity with the agent / MCP ecosystem.
• Ability to thrive in an early-stage environment: write the roadmap as you go, make decisions with incomplete information.
• Collaborative, low-ego approach.

Nice-to-Haves

• None explicitly listed beyond the above production LLM/agent ecosystem experience and domain knowledge.