Foundation Model Data Engineer
Core
Lead the end-to-end strategy, creation, and curation of massive datasets (pre-training, SFT, RLHF) to power next-generation multimodal foundation models.
Role type
Senior IC foundation model data engineer (data-centric AI)
Builds
Pre-training corpora, post-training alignment datasets, and multimodal (vision/video) pipelines for frontier AI models
Domain
AI infrastructure / Large Language Models (LLMs) / Multimodal AI
Deliverable
production ML models
Required skills
Python (high-performance, multiprocessing, multithreading), petabyte-scale data engineering, dataset reconstruction from web crawls, RLHF/DPO pipeline design, data profiling & bias analysis, synthetic data generation, Spark/Ray/WebDataset/Parquet
Preferred skills
Computer vision curation, multimodal crawling, taxonomy design for reasoning/coding/math, research background in data-centric AI
Technologies
Python, Spark, Ray, WebDataset, Parquet
Responsibilities
Define data mix strategies for pre-training; design pipelines for petabyte-scale cleaning and deduplication; lead post-training dataset development (SFT, RLHF/DPO); drive multimodal (video/vision) data acquisition and processing; develop high-throughput ingestion scripts; conduct statistical analysis on training corpora; design synthetic data generation pipelines
Seniority
Senior, hands-on IC