Staff AI Platform Engineer

Responsibilities

• Design and operate the knowledge graph and ontology to capture grocery data relationships
• Build retrieval systems including vector, graph, and structured search to provide model context
• Develop LLM-powered agents involving tool-use, multi-step reasoning, and orchestration
• Build serving infrastructure for agents to ensure reliability and cost-effectiveness
• Establish evaluation sets, LLM-as-judge harnesses, tracing, and observability metrics
• Build data pipelines and experimentation workflows on Databricks and MLflow

Requirements

• 5+ years of experience building production software, data, or ML systems
• Strong systems and API design skills using Python
• Hands-on production experience with LLM systems (RAG, agents, tool-use, prompt engineering)
• Experience with evaluation frameworks and measuring model quality
• Solid data engineering foundations including pipelines and data modeling
• Proficiency with modern cloud data stacks such as Databricks, Spark, or MLflow

Preferred Qualifications

• Experience with knowledge graphs, ontologies, or graph databases
• Proficiency with vector stores like Pinecone, Weaviate, or pgvector
• Familiarity with agent frameworks like LangGraph or MCP protocols
• Experience with MLOps and model serving at scale
• Background in grocery, retail, or complex enterprise data domains

About the Company

Afresh is an AI platform for grocery dedicated to reducing food waste. By leveraging proprietary AI designed for high-volatility environments, Afresh empowers major grocers to make smarter decisions, preventing millions of pounds of food waste annually.