Senior AI Backend Developer

Senior AI Backend Developer Ramat Gan | 3 Hybrid Build and ship production AI agents, and the backend they run against. Our agent platform is built with Python and LangGraph, serves multiple client tenants, and integrates with our .NET backend and real production data. You own agents end-to-end — reasoning, tool calling, state, retrieval, deployment — and build the REST APIs behind them. This is a hands-on engineering role at the intersection of AI agents, backend engineering, and real-world product development. What You'll Do ·; Design and ship multi-step agentic workflows using LangGraph and other agent orchestration patterns · Wire agents to real systems — REST APIs, databases, MCP tool servers · Build RAG pipelines: embeddings, vector search, retrieval quality · Run AI agents in production — streaming, state persistence, human-in-the-loop, tracing, evaluation, and reliability · Design and build scalable REST APIs in C# / .NET for complex web and native mobile apps · Design and optimize SQL Server databases, queries, and stored procedures · Work with product to turn requirements into working agents and services across web and mobile · Use modern AI coding tools to accelerate development while maintaining high engineering standards Requirements — Agentic AI · 2+ years of hands-on experience building and shipping LLM agents or agentic systems in production · Hands-on experience with LangGraph or similar agent orchestration frameworks such as LangChain, AutoGen, CrewAI, or Semantic Kernel · Strong Python skills, including async programming and experience with FastAPI or similar frameworks · Solid on tool/function calling, memory and state, RAG, multi-agent coordination · Shipped against LLM APIs — OpenAI, Anthropic, AWS Bedrock, Azure OpenAI Requirements — Backend · 7+ years C# / .NET (10) building backends for complex web and native mobile apps · REST API design and 3rd-party API integration · 4+ years SQL Server — schema design, queries, stored procedures · Entity Framework Core for data access and modeling · Reads and supports other people's code · Uses AI coding tools (GitHub Copilot, Claude Code) as part of the workflow Advantages · Vector databases (Pinecone, OpenSearch) and embedding pipelines · MCP (Model Context Protocol) · AWS (Bedrock, ECS) or Azure · LLM observability and eval — LangSmith, Langfuse · CI/CD