Member of Technical Staff, Mid-training

About the role

The mid-training team at xAIxAI aims to develop an omniscient model capable of comprehending the universe through various modalities including text, images, videos, and audio. To achieve this, we are seeking experienced engineers specializing in multimodal mid-training data.

Tech Stack

• Python
• JAX and XLA
• SparkRay

Focus

• Design synthetic data algorithms that enhance model intelligence.
• Distill the intelligence of leading models into compact, efficient models using synthetic data generation.
• Optimize mid-training data mixes to elevate model performance.
• Develop robust and varied evaluation methods for mid-training checkpoints.

Ideal Experience

• Expertise in machine learning and large-scale model scaling, with proficiency in various scaling laws.
• Strong experimental design and problem-solving skills.
• Familiarity with advanced techniques for creating AI training data across multiple modalities (text, image, audio, video).
• Proficiency in Spark, Ray, and other large-scale data processing frameworks.

Interview Process

• Initial review of CV and statement of work.
• 15-minute phone interview to assess basic qualifications.
• Four technical interviews:
• Coding assessment in a chosen language.
• Technical sessions (two) focusing on problem formulation and solution design.
• Meet the Team session where candidates present their past work and vision for xAI.
• Process completion within one week.

Benefits

• Annual salary range: $180,000 - $440,000 USD.
• Total rewards package includes base salary, equity, health benefits, retirement plans, insurance, and various perks.

Equal Opportunity Employer

xAIxAI is committed to fostering an inclusive environment and is an equal opportunity employer.

California Consumer Privacy Act (CCPA) Notice

This notice complies with California's CCPA requirements.