Machine Learning Engineer Location: Cupertino, CA The Machine Learning Engineer will join the Channel Strategy and Operations team to build intelligent, multimodal agentic systems focused on troubleshooting, anomaly detection, and operational automation. This role will support two major initiatives: a troubleshooting agent leveraging LLMs/VLMs to analyze databases, images, video, and enterprise signals to identify anomalies, generate reasoning-based insights, and communicate recommendations across stakeholders and downstream agents; and an order approval automation platform that uses image-based analysis to evaluate store readiness, supply conditions, merchandising compliance, and workspace cleanliness.
THE OPPORTUNITY FOR YOU - Develop agentic flows that process visual (images/video) and textual signals to diagnose store-level technical issues (e.g., hardware failures or lighting malfunctions).
- Design and implement RAG pipelines that connect field observations to internal technical training documentation and repair databases.
- Design and execute experiments to validate the accuracy of the agent's diagnostic reasoning and the relevance of its repair suggestions.
- Perform deep-dive analysis on agent failures or hallucinations and translate findings into concrete improvements in the retrieval or reasoning logic.
- Proven track record to apply VLM or Computer Vision Skill to solve a real-world problem in production.
- Own and iterate on systems that capture user feedback on repair success to continuously refine the agent's diagnostic accuracy.
- Turn the need for reduced downtime and faster compliance fixes into practical ML architectures and roadmap milestones.
KEY SUCCESS FACTORS - 3+ years building Machine Learning systems professionally and successfully releasing automated solutions to production
- Proven experience working with Vision Language Models (VLMs) and multimodal AI systems that process and reason across image, video, and text data
- Demonstrated ability to extract insights from media data, including image and video analysis, image-to-text understanding, OCR/text enhancement, and multimodal data interpretation
- Experience building agentic AI systems using frameworks such as Google ADK (preferred) or LangGraph
- Ability to integrate MCP-enabled workflows with enterprise data sources and design collaborative multi-agent architectures capable of reasoning, communicating, and orchestrating actions autonomously
- Strong Python programming skills with proven experience building production-grade AI/ML applications and APIs
26-156824
