Staff, Software Engineer

Walmart · Retail · Hoboken, NJ +1

Staff Software Engineer to design and build next-generation intelligent systems, focusing on deployable, self-improving, agent-driven systems with evaluation frameworks and protocol-based integrations. The role involves building agent orchestration and developer-facing AI systems, emphasizing iteration speed and practical impact. Key responsibilities include developing end-to-end agentic systems, AI-assisted development workflows, agent evaluation frameworks, feedback-driven systems, MCP-based integrations, multi-agent workflows, and introducing observability patterns. The role requires experience building production systems integrating AI/ML, familiarity with LLM-based systems and agent workflows, and hands-on work with evaluation and feedback loops.

What you'd actually do

  1. Design and build end-to-end agentic systems capable of multi-step reasoning, planning, and tool execution.
  2. Develop AI-assisted development workflows that accelerate how software is written, tested, and iterated.
  3. Build agent evaluation frameworks (offline + online) to measure quality, reliability, and system behavior.
  4. Implement feedback-driven systems that continuously improve outputs using real-world signals.
  5. Architect and integrate MCP-based systems to standardize interactions between models, tools, and services.

Skills

Required

  • Python
  • modern backend development
  • building usable systems
  • integrating external tools, APIs, or structured data into intelligent workflows
  • design and ship end-to-end solutions quickly
  • balancing experimentation with reliability
  • working in open-ended problem spaces with evolving requirements
  • ownership mindset

Nice to have

  • AI-powered developer tools or automation systems
  • agent frameworks, orchestration layers, or evaluation harnesses
  • protocol-driven systems (e.g., MCP or similar abstractions)
  • human-AI interaction
  • rapid prototyping
  • unconventional system design
  • building things quickly and iterating based on real feedback

What the JD emphasized

  • production systems
  • end-to-end agentic systems
  • agent evaluation frameworks
  • feedback-driven systems
  • shipping, learning, and improving systems continuously

Other signals

  • design and build next-generation intelligent systems
  • deployable, self-improving systems
  • agent-driven systems
  • evaluation frameworks
  • protocol-based integrations
  • agent orchestration
  • developer-facing AI systems
  • iteration speed
  • practical impact
  • systems that learn from usage
  • end-to-end agentic systems
  • multi-step reasoning, planning, and tool execution
  • AI-assisted development workflows
  • agent evaluation frameworks
  • feedback-driven systems
  • MCP-based systems
  • multi-agent workflows
  • observability, debugging, and traceability in AI-driven systems
  • AI-native application development
  • production systems that integrate AI/ML components
  • LLM-based systems, agent workflows, or tool-augmented models
  • evaluation, feedback loops, or iterative system improvement
  • end-to-end solutions quickly
  • shipping, learning, and improving systems continuously