Staff, Software Engineer - Backend

Walmart Walmart · Retail · Bentonville, AR

Staff Software Engineer to build and maintain GenAI experience platform capabilities, focusing on semantic routing, agent-to-agent communication, and evaluation pipelines for GenAI-powered services.

What you'd actually do

  1. Design and develop platform features enabling advanced semantic routing for GenAI-powered services, ensuring requests are intelligently directed to the most appropriate models or agents based on context and intent.
  2. Build and maintain evaluation pipelines for semantic router data, including the development of metrics, dashboards, and automated tests to assess routing accuracy, system performance, and user satisfaction.
  3. Collaborate with applied researchers and data scientists to continuously improve semantic routing algorithms and integrate the latest advancements in natural language understanding.
  4. Develop and implement agent-to-agent (A2A) communication protocols to enable seamless orchestration and collaboration between multiple GenAI agents and services within the platform.
  5. Contribute to the design and development of platform features using microservices (FastAPI) and event-driven architecture (Kafka, SSE, WebSocket)—all in Python.

Skills

Required

  • Python
  • high-performing, production-quality code
  • Python-based web services
  • FastAPI
  • Git
  • Linux environments
  • containerization technologies
  • Docker
  • 4+ years of industry experience

Nice to have

  • GenAI-based applications
  • large language models
  • generative AI frameworks
  • Hugging Face Transformers
  • LangChain
  • OpenAI API
  • productionize and evaluate GenAI models
  • prompt engineering
  • vector databases
  • Pinecone
  • FAISS
  • Weaviate
  • orchestration tools for GenAI workflows
  • semantic routing
  • analyze, test, and improve semantic routing algorithms
  • responsible AI principles
  • model evaluation
  • bias mitigation
  • production monitoring
  • large-scale data pipelines
  • e-commerce
  • recommender systems
  • MLOps practices and tools
  • MLflow
  • Kubeflow
  • Airflow
  • observability and monitoring tools for production AI systems

What the JD emphasized

  • semantic routing
  • evaluation pipelines
  • natural language understanding
  • agent-to-agent (A2A) communication protocols
  • GenAI-based applications
  • productionize and evaluate GenAI models
  • prompt engineering
  • vector databases
  • semantic routing
  • analyze, test, and improve semantic routing algorithms

Other signals

  • GenAI
  • semantic routing
  • agent orchestration
  • evaluation pipelines