JR AI Agentic Developer

Junior developer (0-3 years) with hands-on experience designing and shipping agentic AI systems. Fluent with AI-assisted development tooling (Cursor, Claude Code) and comfortable building production Python services. Strong command of core agentic concepts - ReAct loops, plan-and-execute, orchestrator-worker patterns, tool/function calling, short- and long-term memory, and multi-agent handoffs - with the judgment to know when an agent is the right architecture and when a deterministic workflow is. Startup experience and supply chain or ERP domain exposure are significant differentiators.

Core Skills (Required)

• Agentic Frameworks: LangChain / LangGraph (state graphs, checkpointing, conditional edges, human-in-the-loop) or CrewAI (role-based crews, task delegation, sequential and hierarchical processes)
• Agentic AI Concepts: ReAct, plan-and-execute, reflection, tool calling and schema design, memory management, context-window strategy, guardrails, and agent evaluation
• AI Development Tooling: Cursor, Claude Code, MCP servers, prompt iteration and debugging
• API Development: FastAPI - async endpoints, Pydantic models, dependency injection, SSE/streaming responses, background tasks; used to expose agent workflows as production services
• Programming: Python (async/await, typing, pytest), Git, Docker
• Location: open to candidates anywhere in the United States; remote, hybrid, or onsite. Must be authorized to work in the U.S.

Preferred Skills

• n8n: workflow orchestration, webhook triggers, scheduled runs, integrating agents with external systems and business tooling
• RAG: chunking strategies, embeddings, hybrid (semantic + keyword) search, re-ranking, citation grounding, retrieval evaluation
• Vector Databases: pgvector, Pinecone, Qdrant, Weaviate, or Chroma
• Graph Databases: Neo4j and Cypher; GraphRAG and knowledge-graph-backed retrieval
• Kubernetes: containerization, Helm charts, deploying and scaling agent services on a cluster
• Startup Experience: early-stage or small-team background - comfortable with ambiguity, owning a feature end to end, moving from prototype to production quickly, and wearing multiple hats without waiting for process
• Supply Chain Domain: exposure to supply chain, procurement, logistics, or ERP systems (Oracle, SAP, NetSuite) - understanding of purchase orders, suppliers, inventory, demand planning, and fulfillment workflows