Data Engineer

xAI•Palo Alto, CA

13h

About The Position

At xAI, we are building AI systems that push the frontier of human knowledge and scientific discovery. High-quality data is fundamental to every stage of that mission. Our Data team is responsible for ensuring that the models are trained on the right data, in the right form, at the right quality, across every phase of the training lifecycle. This includes partnering closely with acquisition teams to identify where valuable data can be sourced, determining what data is needed to improve model performance, and building the production pipelines and systems that transform raw inputs into high-quality training data at scale. We work at the intersection of data, infrastructure, and machine learning to ensure our models train effectively and reliably. As a Data Engineer / AI Engineer on xAI’s Data team, you will be responsible for developing the systems, processes, and production code that power data acquisition, preparation, quality evaluation, and delivery for model training. You will work closely with acquisition teams, ML engineers, and software engineers to identify data needs, build scalable data pipelines, and continuously improve the quality of the data that shapes model behavior. The ideal candidate combines strong software engineering fundamentals and excellent coding practices with deep intuition for statistics, neural networks, and how data quality influences training outcomes.

Requirements

Bachelor’s degree in computer science, data science, physics, mathematics, or a STEM discipline
1+ years of data/software engineering experience (internship experience is applicable)
Experience in implementing or analyzing language models or neural networks

Nice To Haves

Professional experience in analytics, data science, machine learning, or data engineering
Experience building and operating production data pipelines for neural network or large-scale machine learning workloads
Strong experience with Python and the broader ecosystem of libraries and tools used in modern machine learning and data development
Experience working with Parquet or similar columnar storage formats in large-scale data systems
Familiarity with Kubernetes and distributed production environments
Experience developing predictive models and machine learning pipelines, including clustering, forecasting, anomaly detection, or related techniques
Experience working with very large-scale datasets, including terabyte- to petabyte-scale data systems
Strong statistical intuition and the ability to use quantitative analysis to guide technical and product decision, including familiarity of scaling ladder design studies
Ability to operate effectively in a dynamic environment with evolving priorities, changing requirements, and fast-moving technical challenges
Demonstrated ability to take ownership of ambiguous problems, drive projects independently, and develop new expertise where needed

Responsibilities

Analyze the performance and impact of data used throughout the model training lifecycle
Investigate anomalous model behavior and rigorously identify the data issues that drive poor downstream performance
Design, build, and improve the data cleaning, transformation, and quality-control steps required to produce high-quality training data
Research, evaluate, and develop frontier methods for improving data quality and effectiveness in AI model development
Apply statistical techniques and empirical analysis to make informed, data-driven decisions about dataset quality and model outcomes
Partner across teams to identify where data needs exist and define the highest-impact opportunities for new data acquisition and improvement
Build and maintain production-grade data pipelines, tooling, and software systems that ingest, process, validate, and deliver data for training
Develop metrics, evaluation frameworks, and monitoring systems to assess how data quality influences model behavior at scale
Fuse data from multiple sources into reliable, usable datasets for research and production model training
Create shared datasets, tooling, and internal data products that enable other teams to analyze, debug, and improve model performance