Software Development Engineer III - AI

Expedia Expedia · Hospitality · Austin, TX

Software Development Engineer III role focused on designing and implementing AI-powered fraud and abuse defenses, building and operating cloud-native decisioning and automated remediation systems, and applying AI-assisted SDLC practices. The role involves building agentic workflows, simplifying the platform, and ensuring SLOs and observability. Requires experience with Gen AI and agentic systems, distributed cloud-native engineering, and AI coding assistants.

What you'd actually do

  1. Contribute to the design and implementation of low-latency risk decisioning (signals + rules + models) and automated remediation, owning the quality and reliability of the services and components you build.
  2. Apply AI-assisted SDLC practices end to end: AI-accelerated design and code generation, prompt and context management for coding agents, human-in-the-loop review of AI-generated code, automated test generation, and CI/CD guardrails.
  3. Build, deploy, and operate agentic workflows and multi-agent systems — decomposing tasks, orchestrating tools and function calls, managing memory, and embedding human-in-the-loop controls — to automate and reimagine manual operational processes, not just augment them.
  4. Help simplify and modernize the platform (streaming/data pipelines, microservices, CI/CD, configuration-driven controls) so that teams can iterate quickly and safely.
  5. Implement and maintain SLOs, observability, and safety/rollback mechanisms for the services you own, ensuring security, privacy, and compliance requirements are met by design.

Skills

Required

  • 5+ years of software engineering
  • building and operating high-scale production backend services
  • shipping Gen AI and agentic systems to production
  • system design (LLD)
  • API design
  • data modeling for AI-enabled services
  • distributed cloud-native engineering at scale (AWS, GCP, or Azure)
  • microservices
  • API-driven design
  • SQL/NoSQL databases
  • data streaming/processing (Kafka, Flink, Spark)
  • Hands-on experience using AI coding assistants and agents (e.g., Claude Code, Cursor, GitHub Copilot, or equivalent)
  • applying AI-assisted SDLC practices across design, implementation, reviews, and test generation
  • strong prompt/context management
  • human-in-the-loop review of AI-generated code
  • Experience building, monitoring, and debugging LLM and multi-agent applications
  • frameworks and platforms such as LangChain, LangGraph, Langfuse, or equivalent
  • RAG-based architecture experience
  • data orchestration frameworks such as LlamaIndex
  • vector databases such as Pinecone, or equivalent
  • Exposure to various LLM providers such as OpenAI, Gemini, and Anthropic

Nice to have

  • Track record automating and reimagining manual operational processes with agen

What the JD emphasized

  • shipping Gen AI and agentic systems to production
  • AI-assisted SDLC practices
  • agentic workflows and multi-agent systems
  • human-in-the-loop review of AI-generated code

Other signals

  • AI-powered fraud and abuse defenses
  • agentic workflows and multi-agent systems
  • AI-assisted SDLC practices