Senior Machine Learning Data Pipeline Engineer

NVIDIA•Santa Clara, CA

About The Position

NVIDIA is hiring a Senior Machine Learning Data Pipeline Engineer to invent and build next-generation data pipelines for end-to-end autonomous driving foundation models. The role develops scalable pipelines and agentic workflows that help DNN teams identify data gaps and discover, prepare, and deliver appropriate training data. It converts model limitations and data needs into reusable engineering solutions by integrating data infrastructure with LLMs, vision-language models, and retrieval techniques to automate data solutions.

Requirements

Bachelor’s or master’s degree in Computer Science, Mathematics, Electrical Engineering, or a related quantitative field, or equivalent experience.
At least 5 years of experience in software or data engineering.
Strong Python and SQL skills demonstrated through production services, APIs, or data pipelines.
Hands-on development and optimization of data solutions, including diagnosing DNN model performance and reliability issues.
Ability to develop and analyze automatic workflows connecting LLMs to tools and data services, including tool calling and workflow state management.
Solid understanding of deep learning, embeddings, and LLM/VLM inference, including how training datasets are prepared and consumed by DNN teams.
Experience delivering data solutions from unclear requirements through production, including setting acceptance criteria with DNN engineers and data specialists and confirming outcomes.

Nice To Haves

Hands-on work on foundation model data and training pipelines, including smaller-scale models built from scratch.
A track record of shipping LLM/VLM-based mining, multimodal retrieval, or curation services.
Background in autonomous driving or robotics data, including temporal video, sensor metadata, scenario definitions, or model failure analysis.

Responsibilities

Design, build, and own scalable data pipelines that automate data discovery, selection, and curation to meet DNN training data needs.
Develop agentic workflows that proactively identify model data needs, plan and implement mining tasks using data tools and model inference, and refine mining strategies through human review and downstream model feedback.
Deploy and operate pipelines and workflows as reliable production systems, including workflow orchestration, monitoring, and failure recovery.
Partner with MLEs and data analysts to turn recurring data needs and model failure cases into reusable mining capabilities, define acceptance criteria, and improve scenario coverage through feedback.