Member of Technical Staff, Pretraining Data (AI)

Location: San Francisco, United States; onsite

Base salary: $300,000–$400,000 USD, plus meaningful equity grant depending on experience and background.

Company

hirify.global is a tech-first frontier AI research company building its own models end-to-end to accelerate science and technology.

What you will do

• Build pretraining data pipelines for repository-scale code and technical documents.
• Improve public data through extraction, filtering, deduplication, and processing of web pages, open-source code, and public datasets.
• Decide data mixtures for training runs and design ablations and data-scaling experiments to justify them.
• Direct agents to run experiments and pipelines in parallel, then evaluate which results are trustworthy.
• Extend code-first data mixtures into scientific domains where models underperform.
• Maintain corpus provenance, decontamination against evaluations, and reproducible versions.

Requirements

• Shipped a data change into a large pretraining run and demonstrated the ablation that justified it.
• Built data pipelines at the scale of trillions of tokens using Spark, Ray, Beam, or similar technologies, with strong Python skills.
• Strong enough knowledge of PyTorch or JAX to run independent training experiments.
• Hands-on experience with at least two of deduplication, quality classifiers, mixture optimization, data scaling laws, long-context data, or code data.
• Ability to distinguish real experimental results and explain the reasoning clearly in writing.

Culture & Benefits

• Full-time onsite work in San Francisco.
• Work closely with researchers responsible for model architecture and training runs.
• Use agents to execute engineering, ablation, pipeline, and analysis work in parallel.
• Competitive benefits and a meaningful equity grant.