Job Summary:
The Scotts Miracle-Gro Company is a leader in the gardening industry, and they are seeking an Agent Engineer to own the AI agent development lifecycle. This role involves designing and delivering production-grade AI agents to solve business challenges across customer service and internal workflows.
Responsibilities:
• Own end-to-end agent development — from requirements and design through pilot, deployment, monitoring, and iterative improvement.
• Build production-grade AI agents using Sierra and complementary tools (LangChain, ADK, or similar orchestration frameworks) for customer-facing and internal use cases.
• Design agent architectures: conversation flows, tool use, RAG integration, human handoff logic, and multi-channel deployment (chat, SMS, voice, email).
• Implement prompt engineering, guardrails, and evaluation pipelines to ensure agents meet quality, safety, and brand standards.
• Become the in-house expert on Sierra and related agent platforms — learn the Agent SDK, Agent Studio, and deployment options to maximize platform value.
• Evaluate when to use platform-native capabilities vs. custom extensions; recommend build vs. buy vs. configure for new agent use cases.
• Integrate agents with SMG systems: CRM, e-commerce, knowledge bases, and internal tools. Ensure secure, compliant access to data and APIs.
• Establish reusable patterns, templates, and best practices for agent development that the CoE and FDEs can leverage.
• Partner with Digital and CS teams to identify high-value agent opportunities and define success metrics.
• Track agent performance (resolution rates, satisfaction, deflection, latency) and drive continuous improvement based on data and user feedback.
• Document agent designs, runbooks, and operational procedures. Support responsible AI practices and governance requirements.
Qualifications:
Required:
• Bachelor's or Master's degree in Computer Science, Software Engineering, or a related technical field.
• 3+ years of software engineering experience, with 1+ years building or integrating AI/LLM-powered applications or agents.
• Strong programming fundamentals — Python required; experience with React, TypeScript/JavaScript and API design is a plus.
• Hands-on experience with generative AI: LLMs (GPT-4, Claude, Gemini), prompt engineering, RAG, vector databases, and agent orchestration (LangChain, Flowise, or similar).
• Ability to learn new platforms quickly — you will ramp on Sierra (or equivalent) and become productive within weeks.
• Track record of shipping production software with clear ownership of design, implementation, and iteration.
• Clear communicator who can explain technical decisions to non-technical stakeholders and align on priorities.
Preferred:
• Experience with Sierra, Voice AI, or other enterprise conversational AI platforms.
• Familiarity with customer service, e-commerce, or CPG/retail domains.
• Experience with cloud platforms (GCP, AWS, or Azure) and CI/CD for ML or agent deployments.
• Exposure to agent evaluation, A/B testing, or observability for AI systems.
• Understanding of responsible AI, data governance, and compliance considerations for customer-facing AI.
Company:
Founded in 1868 in Marysville, Ohio, The Scotts Miracle-Gro Company is the leading marketer of branded consumer lawn and garden products in North America. Founded in 1868, the company is headquartered in Marysville, USA, with a team of 1001-5000 employees. The company is currently Late Stage.