We are looking for a skilled Agentic AI Developer to build and productionize Vertex AI RAG systems, design tool-using agents, and integrate with vector and graph databases. You will own end-to-end workflows from ingestion → retrieval → agent orchestration → evaluation → deployment.
Roles and Responsibilities
- Design and implement RAG pipelines on Vertex AI: chunking, embeddings, indexing, retrieval, reranking, and grounding.
- Build agentic workflows using Python-first frameworks (tool use, planning, reflection, structured outputs).
- Integrate with Vector DBs (Vertex Vector Search, Pinecone, Weaviate) and Graph DBs (Neo4j, JanusGraph, Neptune).
- Develop data ingestion/ETL pipelines from PDFs, documents, web sources, and internal systems.
- Define and execute evaluation strategies for retrieval quality, hallucination checks, and system improvement.
- Ship production-grade solutions: APIs, monitoring, CI/CD, performance optimization, and security best practices.
Required Skills
- Strong Python (async, testing, typing, packaging) and engineering practices.
- Proven experience with RAG solutions, embeddings, hybrid search, and prompt/schema design.
- Hands-on experience with Vertex AI/GCP (IAM, monitoring, Cloud Run/GKE, storage).
- Experience with at least one agentic framework (LangChain, LangGraph, LlamaIndex, Semantic Kernel, AutoGen).
- Knowledge of vector search and at least one vector DB in production.
- Familiarity with graph data modeling and basic querying (Cypher, Gremlin, SPARQL).
Nice-to-Have: Knowledge graphs, streaming/messaging (Pub/Sub, Kafka), multilingual retrieval, evaluation tooling, or frontend integration (React/Next.js).