The NVIDIA System-On-Chip(SOC) Design team is looking for a top AI Engineer with curiosity about SOC design automation, RTL integration, and chip build and assembly now. If you are interested in using AI to upgrade the conventional SOC Design flow, come and join us. We need you to be passionate about AI+Hardware. You are expected to help us to build AI application-layer services which would boost HW execution team's work efficiency, includes : assistants, retrieval and Q&A; workflow automation; and develop AI agent for SOC Design-related tasks. You will be shipping and operating AI services (APIs, orchestration, RAG, evaluation), evaluating and using modern frameworks and tools, such as LangChain (and similar stacks), RAG pipelines, and coding-agent / IDE-centric workflows (e.g. Claude Code-class assistants, reusable skills / playbooks for agents).
What you’ll be doing
- Design, implement, and operate LLM-backed services: APIs, async jobs, streaming responses, and integration with internal tools and data sources.
- Build RAG and knowledge systems: chunking, embeddings, vector retrieval, reranking, access control, and quality/latency tuning.
- Apply agent and orchestration patterns with frameworks like LangChain (or comparable): tool use, multi-step plans, memory, and guardrails—aligned with how SOC Hardware team works.
- Improve developer and engineer experience with AI-assisted coding and repeatable “skills”: prompts, procedures, and small utilities that teams can run consistently (including patterns like Claude Code + structured skills).
- Own reliability and perform evaluation: logging, tracing, regression tests for prompts/pipelines, and metrics for usefulness and safety on proprietary data.
- Co-work with Hardware engineers from Methodology, CAD, and Design teams to scope the problem, propose the solution, implementation (in multiple iterations), and online production-ready features.
What we need to see
- MS/PhD in CS, CE, EE
- 1–5 years of professional experience with a clear focus on AI application / AI service development (building products on top of LLMs, not only ad-hoc scripts).
- Strong Python and experience shipping services (REST/gRPC, containers, basic cloud or on-prem deployment patterns as applicable).
- Hands-on use of LLM application frameworks (e.g. LangChain or equivalent) and RAG (vector DBs, retrieval design, evaluation).
- Familiarity with coding agents and IDE workflows (e.g. Claude Code-style usage) and frameworks (skills, templates, or internal “agent packs”).
- Solid software engineering habits: dependency management, configuration, testing, and clear interfaces for other teams.
- Excellent communication and ability to work with partners who are not AI specialists.
Ways to stand out from the crowd
- Hardware knowledge: RTL Coding capability; Makefile Coding capability; SOC Design know-how; Physical Design know-how; etc—enough to understand user context and data (no requirement to be a chip designer).
- Web development: lightweight UIs, internal portals, or full-stack slices (e.g. React/TypeScript, FastAPI + frontend) for AI features.