Lead Software Engineer, Full Stack (enterprise Platforms Technology)

Capital One Capital One · Banking · McLean, VA +2

Lead Software Engineer to build an AI-powered marketing content layer, integrating LLMs into content generation, channel-specific builds, and AI-assisted compliance workflows. The role involves developing multi-agent orchestration systems, evaluation/observability tooling, and guardrail frameworks for LLM outputs, with a focus on enterprise AI applications.

What you'd actually do

  1. Build software solutions to iterate on AI content generation services - integration of LLM’s into content authoring workflows with prompt management, output validation and Human-In-The Loop review.
  2. Implement AI-assisted compliance review workflows – build classifiers and scoring pipeline that flag possible UDAAP, CAN-SPAM, claims and disclosures before content approval.
  3. Contribute to multi-agent orchestration systems that coordinate content compliance review, audience selection, and campaign dispatch.
  4. Build evaluation and observability tooling for LLM outputs – golden dataset benchmarks, A/B content quality metrics and drift detection.
  5. Define guardrail frameworks for LLM outputs – toxicity filters, brand voice enforcement, compliance boundary checks – and establish operational runbooks.

Skills

Required

  • Python
  • cloud computing (AWS, Microsoft Azure, Google Cloud)
  • LLM integration
  • model API
  • prompt engineering
  • AI-augmented feature generation
  • Docker
  • Kubernetes
  • AWS tools and services

Nice to have

  • Java
  • Go
  • TypeScript/JavaScript
  • Node.js
  • Marketing Technology
  • ESP’s
  • campaign management
  • content personalization
  • digital marketing platforms
  • MJML
  • Apache Velocity

What the JD emphasized

  • AI-assisted compliance workflows
  • AI-assisted compliance review workflows
  • guardrail frameworks for LLM outputs

Other signals

  • LLM-integrated content generation pipelines
  • AI-assisted compliance workflows
  • multi-agent orchestration systems
  • evaluation and observability tooling for LLM outputs
  • guardrail frameworks for LLM outputs