Stay hands-on: design and ship the code, methods, and reference architectures that bring RAG, inference, and multi-agent, long-horizon workflows to life on our stack (NeMo, Nemotron, NeMo Agent Toolkit, NIM, Dynamo, TensorRT-LLM) and open tools (vLLM, LangChain, vector DBs, MCP, A2A). Build breadth across the agentic AI lifecycle, with depth in a few areas that fit you: fine-tuning (PEFT, SFT), post-training and RL from verifiable rewards, reasoning, advanced RAG, multi-agent workflows, skills/harness engineering, agent evaluation and observability, and production inference.