Machine Learning Engineer, GeForce G-Assist

What you'll do

• Evaluate SLM behavior and mitigate conversation or context contamination, including state drift, prompt leakage, and retrieval cross-talk.
• Optimize local inference with llama.cpp and performance-sensitive C/C++ code; design RAG systems and support agentic workflows.
• Collaborate on text and multimodal model systems, including hybrid architectures combining local SLMs with cloud-based models.

What they ask for

• Strong C/C++ skills in systems-level or performance-sensitive environments and proficiency in Python.
• Hands-on experience with llama.cpp or similar local inference frameworks.
• Hands-on SLM evaluation experience, including task-based and conversational testing; understanding of conversation dynamics, long-context behavior, and contamination challenges.
• Knowledge of SLM and VLM architectures, retrieval technologies and language-model integration, and agentic AI patterns such as tool use and planning.

Nice to have

• Experience contributing to language or multimodal models for user-facing products, features, or workflows.
• Experience collaborating across product, platform, or systems teams and translating user needs into measurable improvements.

Pay

152,000 USD - 241,500 USD (base salary, Level 3); 184,000 USD - 287,500 USD (base salary, Level 4)

Tools

Small Language Models (SLMs), VLM, llama.cpp, C/C++, Python, retrieval-augmented generation (RAG)

AI focus

Build and improve an on-device AI assistant using SLMs, multimodal models, retrieval, hybrid cloud architectures, and agentic workflows; focus on model reliability and local inference performance.